QUANTITATIVE SEPARATION OF LITHIUM. 17 



tions to which were added none, 0.0236 and 0.0118 o;ram, respectively. 

 The sohitions contained also 0.005 gram of silica, 0.0006 gram of 

 ferric oxid and alumina, 0.0232 gram of calcium oxid, and 0.0254 

 gi'am of magnesium oxid. In view of the general opinion of the 

 uncertainty of the phosphate method and the possibility of loss of 

 lithium with the various precipitates and in the removal of ammonium 

 salts by heat, it seems unlikely that many will attempt to determine 

 lithium in this final magnesium filtrate. 



^luEMAN ^ found the i)hosphate method unreliable by reason of 

 impurities with the ])hosphate ])recipitate which made it too high, 

 but which when removed left too little lithium phosphate. In an 

 effort to obtain a suitable organic solvent for the separation of 

 lithium chlorid, he tried water-free pyridin, anilin, methylanilin, 

 o-toluidin, and phenol. Of these pyridin was the only one to dis- 

 solve the lithium chlorid well. Anilin dissolved it a little, and the 

 others did not dissolve it at all. He gave as a method of separation 

 the treatment recommended by Kahlenberg and Krauskopf.^ He 

 filtered off and weighed the material which was insoluble after the 

 ignition of the sulphate and which he considered ferric oxid and 

 magnesium oxid and subtracted the weight from that of the sul- 

 phates. The insoluble material obtained when the sulphates are dis- 

 solved has usualh^ been considered as consisting in part of silica, and 

 often sufficient magnesium has been found to precipitate as phos- 

 phate from the sulphate solution. Murman stated later ^ that he 

 had not read of the work of Kahlenberg and Krauskopf until after 

 the publication of his paper. 



METHOD USED IX THE BUREAU OF CHEMISTRY. 



Wlien the lithium is to be determined in the same portion of w^ater 

 as the other alkalis and the sulphate, the silica is first removed by 

 two evaporations with hydrochloric acid. The filtrate and washings 

 from the silica determination are concentrated to from 150 to 200 cc 

 and the sulphate determined as barium sulphate. In cases where 

 much sulphate is present it is precipitated according to the methodi 

 of Johnston and Adams.* The occlusion of alkali salts by barium 

 sulphate has been noted by many writers and carefully studied by 

 Allen and Johnston ^ and Johnston and Adams. Their results wero 

 obtained on solutions containing quantities of alkaline sulphates and 

 chlorids much larger than are usually met with in water analysis, 

 On account of the difficulty of obtaining the chlorids in condition to 

 be weighed it has been found expedient to evaporate such a quantity 

 of water that not more than 0.3 to 0.5 gram of mixed chlorids aro 



> Zts. anal. Chem., 1911, 60: 171-174. < J. Amer. Chem. Soc, 1911, SS: 829-845. 



2 Loc. cit. 6 J. Amer. Chem. Soc, 1910, Sg: 588-617, 



3 Zts. anal. Chem., 1911, HO: 27.3. 



40061°— Bull. 153—12 3 



