G. H. Wa?*ren — Niobium and Tantalum. 521 



Weight in. grams : 



Wgt. of 

 Of Nb 2 5 — Ta 2 5 oxides 



(about 3:1) taken Of Ti0 2 extracted Excess 



No. 1 -2500 "2359 '3046 '0687 



No. 2 -2103 '2039 '2580 -0541 



No. 3 '3031 -2698 '2883* '0185 



No. 4 -0178 (Nb Q 5 only) '3605 -0010 niobium 



undissolved 



From No. 4 it would appear that as much as 5 per cent of 

 niobium may pass into solution with an excess of titanium and 

 thus be practically lost in the course of analysis. Although fur- 

 ther study of the most favorable conditions of fusion and solu- 

 tion might be of interest in themselves, the method appears to 

 offer little chance of being modified so as to give more than 

 roughly approximate results. 



The method proposed by T. B. Osbornef was next examined. 

 This depends on the titration of a solution containing the tita- 

 nium and niobium in the lower state of oxidation with potas- 

 sium permanganate, thus oxidizing these elements to the 

 higher oxides, and the subsequent estimation of titanium colori- 

 metrically, while any tantalum, which is not reduced to the 

 lower oxide, present in the original sample, is found by dif- 

 ference. The method of procedure is briefly this : the mixture 

 of the three acids, tantalic, niobic and titanic, is dissolved in 

 hydrofluoric acid and the excess of acid expelled on the water 

 bath. The fluorides thus obtained are dissolved in concentrated 

 hydrochloric acid washed with the same acid into a 100 cc flash 

 (total volume about 50 cc ) and reduced for three-quarters of an 

 hour with amalgamated zinc and a piece of platinum foil in 

 an atmosphere of carbon dioxide at a temperature of 80° C. 

 The reduced solution is cooled thoroughly, diluted with freshly 

 boiled, cold water to about 350 cc and titrated with permanganate. 

 To this solution ammonia is added in slight excess, the pre- 

 cipitate formed, just dissolved in sulphuric acid, and the 

 volume made up to exactly 500 cc . The titanium is then esti- 

 mated colorimetrically, with hydrogen peroxide in 50 cc por- 

 tions of this solution. 



For any except small amounts of titanium the colorimetric 

 method is open to the serious objection that any error in esti- 

 mating the amount of titanium in the aliquot portion (and that 

 error can hardly fail to be considerable where a large amount 

 of titanium is present) is multiplied by ten in estimating the 

 total amount present. Mr. Osborne gives the result of only a 

 single application of the method. The figures are as follows : 



* Still gave a strong reaction for titanium. 



f The Quantitative Estimation of Niobium, this Journal (3), xxx, pp. 328- 

 337, 1885. 



