416 Gooch and Havens — /Separation of 



Art. LIX. — A method for the Separation of Aluminum 

 from Iron ; by F. A. Gooch and F. S. Havens. 



[Contributions from the Kent Chemical Laboratory of Yale University. — LXI.} 



Of the well-known methods for the separation of aluminum 

 from iron — by the action, for example, of an alkaline hydrox- 

 ide in aqueous solution or by fusion of the mixed oxide in 

 potassium or sodium hydroxide ; by reduction of the iron oxide 

 to the metal by heating in hydrogen, with the subsequent solu- 

 tion of the metallic iron in hydrochloric acid ; by boiling the 

 nearly neutral solution of the salts of aluminum and iron with 

 sodium thiosulphate either with or without sodium phosphate ;. 

 by acting with hydrogen sulphide or ammonium sulphide upon 

 solutions of the salts containing also an ammoniacal citrate or 

 tartrate — no single process can be said to be ideal as regards 

 directness, rapidity and accuracy of working. We have deemed 

 it not superfluous, therefore, to attempt the utilization of a 

 reaction which should apparently be capable of effecting directly 

 and quickly the separation of aluminum from iron under con- 

 ditions easily attainable. 



It is known* that the hydrous aluminum chloride AlClg.GH^O 

 is very slightly soluble in strong hydrochloric acid, while ferric 

 chloride, on the other hand, is extremely soluble in that 

 medium. It is this difference of relation of which we wished 

 to take advantage. 



It appeared at the outset that crude aluminum chloride 

 could be freed from every trace of a ferric salt by dissolving it 

 in the least possible amount of water,' saturating the cooled 

 solution with gaseous hydrochloric acid, filtering upon asbestos 

 in a filtering crucible or cone, and washing the crystalline pre- 

 cipitate with the strongest hydrochloric acid. Aluminum 

 chloride prepared in this way gave no trace of color when dis- 

 solved in water and tested with potassium sulphocyanide. The 

 correlative question as to how much aluminum chloride goes 

 into solution under the conditions was settled by taking a por- 

 tion of the pure aluminum chloride, dissolving it in a very 

 little water, diluting the solution' with strong hydrochloric 

 acid, saturating the cooled liquid with the gaseous acid, filter- 

 ing on asbestos, precipitating by ammonia the aluminum salt 

 in the filtrate and weighing the ignited oxide. 



From lOcm^ of such a filtrate we obtained in two deter- 

 minations 0-0022 grm. and 0*0024 grm. of the oxide, the mean 

 of which corresponds to 23 parts of the oxide or 109 parts of 

 the hydrous chloride in 100,000 parts of the strong hydrochloric 

 acid. This degree of solubility, though inconsiderable when 



* Gladysz : Ber. d. d. chem. Gesell., xvi, 44*7. 



