330 T. B. Osborne — Quantitative determination of Niobium. 



crystallization be several times repeated, as is usually neces- 

 sary, a very considerable loss must occur. Moreover, the 

 tantalum potassium fluoride cannot be entirely freed from nio- 

 bium in this way, for I found in preparing TaF B 2KF that the 

 niobium cannot be entirely removed by simply crystallizing 

 and washing, as I was compelled to recrystallize several times 

 before the TaF 6 2KF gave no evidence of the presence of nio- 

 bium, when tested by reducing with zinc and hydrochloric acid. 

 Titanium will, when present in any considerable quantity, 

 render the application of this method still more uncertain on 

 account of the intermediate solubility of the potassium titanium 

 fluoride and its close resemblance to the NbOF 3 2KF. 



From this description it will readily be seen that this method 

 is open to serious objection and can only find acceptance in 

 want of a better. I therefore undertook an investigation of 

 the reduction with zinc in acid solutions containing niobium 

 with the hope of discovering a more accurate and less difficult 

 method of determination. The salts of tantalum and niobium 

 I prepared from columbite from Branchville, Connecticut, in 

 the following way : The mineral was ground to pass a sieve of 

 100 meshes to the inch, and 500 grams fused with bisulphate 

 of potassium in a platinum dish, 100 grams at a time. When 

 thoroughly fused the mass was immediately poured into cold 

 water which caused the separated tantalic and niobic acids to 

 granulate and expose a large surface to the action of the water, 

 whereby the solution of the bisulphate of potassium was very 

 materially promoted and the lumps broke up easily on stirring, 

 the tantalic and niobic acids separating partly, in a flocculent 

 and partly in a granular form. 



After washing by decantation the residue was treated with 

 ammonium sulphide and then filtered and washed to remove 

 any tin or tungsten which might be present. Hydrochloric 

 acid was then added and after washing, the residue was ob- 

 tained free from iron and manganese. The residue was next 

 treated with hydrofluoric and sulphuric acids, in order to remove 

 any silica, precipitated with ammonia and washed free from sul- 

 phates. The tantalic and niobic acids thus obtained were dis- 

 solved in an excess of hot hydrofluoric acid and carbonate of 

 potassium added gradually, and after cooling, the TaF 6 2KF 

 filtered off. The crude TaF 6 2KF was recrystallized several 

 times and washed till no reduction whatever took place on 

 testing with zinc and hydrochloric acid. The mother liquor 

 from the crude TaF 6 2KF was then nearly neutralized with 

 carbonate of potassium which caused the solution to nearly 

 solidify from the formation of JSTbOF B 2KF, This product 

 was then recrystallized several times and tested for tantalum 

 by boiling with pure water, when, according to Marignac, 

 if the slightest trace of tantalum were present, it would sep- 



