Extraction of Niobium

The method of extraction of niobium and tantalum compounds from their natural ores does not differ from the process followed in the quantitative examination of the ores, and consists, briefly, in fusing the material with an alkali or alkali salt, extracting the fused mixture of niobates and tantalates with water, and hydrolysing the solution by boiling, whereupon a comparatively insoluble mixture of niobic and tantalic acids or their anhydrides is obtained, which yields the pentoxides, Nb₂O₅ and Ta₂O₅, on being ignited. Only those metals which give rise to acid oxides demand special separation.

The more detailed description of the extraction is conveniently divided into three stages:

1. Preparation of a mixture free from tin, antimony, iron, manganese, etc.
2. Removal of titanium from the mixture.
3. Separation of niobium from tantalum in the product.

The finely powdered niobite or tantalite, which should contain as little titanium as possible, is fused for several hours with a large excess of potassium hydrogen sulphate or sodium hydrogen sulphate in a silica or platinum crucible; the cooled mass is thoroughly extracted by boiling with water, and the precipitate, which consists mainly of niobic and tantalic acids together with some of the sulphates, is digested with ammonium sulphide to remove tin, antimony and tungsten, and to convert any iron or manganese into sulphide, which is removed with hydrochloric acid. Some of the niobic and tantalic acids may also be dissolved by the hydrochloric acid, however. Separation of silica is effected in the usual way by evaporation of the hydrofluoric acid solution of the residue in a platinum dish with addition of sulphuric acid. In order to remove silica without loss of niobium and tantalum through possible vaporisation of the pentafluorides, NbF₅ and TaF₅, extraction of the niobic and tantalic acids with caustic soda or caustic potash has been recommended.

Treatment of the fused ore with ammonium sulphide, as described above for the removal of tin, antimony and tungsten, does not proceed quantitatively; better separation is claimed to result on fusing the mixed niobic and tantalic acids with a large excess of a mixture of sodium carbonate and sulphur and then extracting with water, several refusions being necessary. An alternative method for the removal of tungsten consists in boiling the alkaline aqueous extract of the fused ore with ammonium nitrate, the mixed niobic and tantalic acids being precipitated, while ammonium tungstate is left in solution. In a more recent method the mixed niobic and tantalic acids are fused with potassium carbonate and the aqueous extract treated with sodium chloride; the mixed acids are thereupon precipitated, the tungsten being left in the filtrate as tungstic acid.

Instead of lixiviating with water, the pyrosulphate fusion is followed in a recent process by extraction with tartaric acid solution; the insoluble residue contains silica, tin, and lead, and the solution, after being saturated with hydrogen sulphide for the precipitation of copper, antimony, etc., contains the hydroxides of niobium and tantalum as well as tungsten, titanium, zirconium, rare earth metals, etc.

In addition to potassium hydrogen sulphate and sodium hydrogen sulphate for opening up the ore, potassium carbonate, sodium peroxide, and alkali hydroxides have been employed. The use of potassium hydroxide is preferred in the case of a high-grade ore of low titanium content; it has the advantage over sodium hydroxide that potassium tantalates are soluble in solutions which contain excess of the alkali, whereas sodium tantalates are insoluble.For minerals in which the titanium content is high it has been found preferable to attack the ore with potassium hydrogen fluoride, KHF₂, or concentrated hydrofluoric acid. In one such process the powdered niobite is evaporated almost to dryness with a solution of potassium hydrogen fluoride, and the residue fused and dissolved in hydrofluoric acid, from which crystals of potassium tantalum fluoride, K₂TaF₇, and of potassium niobium oxyfluoride, K₂NbOF₅.H₂O, are obtained on evaporating and cooling. These are freed from iron and manganese either by recrystallising, or better, by previous treatment in solution with hydrogen sulphide.

Older methods of opening up the ore, now only of historical interest, consisted in heating it for several hours with a mixture of sugar charcoal and sodium carbonate in a carbon crucible, whereupon the niobium, tantalum and titanium formed their carbides and nitrides. The product was treated with boiling concentrated hydrochloric acid and dilute hydrofluoric acid, which removed the tin, iron, calcium, and some of the yttrium. The dried residue was carefully heated in a stream of chlorine; the more volatile chlorides of titanium and silicon were thereby removed, and there was left in the tube a mixture of the chlorides of niobium and tantalum, together with small proportions of ferric chloride and tungsten oxychloride. The product was extracted with dilute hydrochloric acid, and the niobium and tantalum chlorides hydrolysed by boiling water to the pentoxides. Moissan obtained a mixture of niobic and tantalic acids very conveniently by heating a powdered niobite with sugar charcoal in the electric oven. Most of the manganese and the greater portion of the iron and silicon were volatilised; the residue, which consisted of the carbides of niobium and tantalum, was dissolved in hydrofluoric acid to which a small quantity of nitric acid had been added. The iron was removed with ammonium sulphide, and potassium fluoride added, whereupon concentration and cooling gave a mixture of potassium tantalum fluoride, K₂TaF₇, and potassium niobium oxyfluoride, K₂NbOF₅.H₂O.

Zirconium can be removed from the mixed precipitated acids by fusing them with potassium carbonate and extracting the melt with cold water. The niobium and tantalum pass into solution as niobate and tantalate of potassium respectively, while the zirconium remains undissolved as the dioxide, ZrO₂. The method is more suited for the removal of zirconium from niobates than from tantalates.

The removal of titanium from mixed niobic and tantalic acids is a difficult matter. Although titanium and niobium compounds display considerable differences in their general behaviour, when the two elements occur together they appear to undergo a change, in consequence of which they become difficult to separate. Niobic acid, for instance, is precipitated from a much more concentrated boiling sulphuric acid solution than is titanic acid; but when the two acids are dissolved together in sulphuric acid the precipitation of the niobic acid does not take place unless the solution has been diluted considerably, i.e. the hydrolysis of the niobium salt is impeded by the presence of titanium. (On the other hand, the hydrolysis of the niobium salt is accelerated by the presence of a small quantity of tantalum.) Again, when titanic acid is fused with potassium carbonate and the melt is extracted with boiling water, only about 1 per cent, of the titanic acid is dissolved; in the presence of niobic acid, however, the greater proportion of the titanic acid passes into solution on the same treatment, and some of the niobic acid remains in the residue. These modifications in the properties of niobium in the presence of titanium were in part responsible for the erroneous assumption of the existence of various new elements in niobium and tantalum ores. More recently it has also been shown that the solubilities of niobic acid, tantalic acid and titanic acid in acidified hydrogen peroxide solution, are affected by the presence of each other, according to the conditions.

A completely satisfactory process for the quantitative separation of titanium from niobium and tantalum has not yet been evolved; nearly all the methods that have been suggested from time to time have subsequently received adverse criticism. One that suffers least in this respect consists in boiling the mixed precipitated acids for several hours with excess of a dilute solution of salicylic acid. The titanic acid is dissolved; the residue is ignited and the process repeated several times, when all the niobium and tantalum are contained in the residue, while the salicylate filtrates contain all the titanium, which is subsequently precipitated with ammonia and estimated as titanium dioxide, TiO₂. Another process of separation, which is stated to be available in the presence of appreciable amounts of titanium, consists in fusing the mixed acids (after removal of tin, antimony, tungsten) with sodium nitrate, or with an alkali and an oxidising agent, which is claimed to prevent the formation of soluble compounds of titanium with niobium or tantalum. After extracting the fused product with water only a very little titanium remains in solution, and this is removed by hydrogen sulphide. The filtrate is boiled until free from hydrogen sulphide, acidified with sulphuric acid, and boiled with sulphurous acid, which precipitates only niobic and tantalic acids. Alternatively, the mixed precipitated acids may be dissolved in acidified hydrogen peroxide and boiled with sulphurous acid; the precipitate contains only a small amount of titanium, and repetition of the process is stated to yield a titanium-free product. Fusion of the mixed acids with a mixture of caustic potash and potassium cyanide has also been recommended for the quantitative separation of titanium; the melt is extracted with hot water, all the titanium remaining in the residue.

In 1905 Hall and Smith investigated all the then known methods for the removal of titanium, and tried various other processes; they were unable, however, to improve on Marignac's method of fractional recrystallisation of the double potassium fluorides. This method has the disadvantage that in the case of the niobium salt protracted and tedious repetition is necessary before it is obtained free from titanium, and the method becomes impossible with small quantities of material.

The close similarity in the chemical behaviour of the compounds of these two elements has rendered their separation extremely difficult. Although many processes have been investigated, the method most in use appears to be that evolved by Marignac as long ago as 1866, or a modification of it. This depends, firstly, on the difference in the solubilities of potassium niobium oxyfluoride, K₂NbOF₅.H₂O (1 part in 12 to 13 parts of water between 17° and 21° C.), and potassium tantalum fluoride, K₂TaF₇ (1 part in 150 to 160 parts of water containing a small quantity of hydrofluoric acid, at the same temperature); and, secondly, on the fact that these two compounds are not isomorphous, and mixed crystals or solid solutions are therefore not produced. For the separation, the mixture of niobic acid and tantalic acid is dissolved in concentrated hydrofluoric acid, potassium fluoride is added in correct quantity, and the whole is carefully concentrated. The double potassium tantalum fluoride is first precipitated in acicular, rhombic needles; the filtrate, on being concentrated, with further addition of hydrofluoric acid and potassium fluoride, yields white, semi-transparent, granular plates of potassium niobium oxyfluoride mixed with the needles of potassium tantalum fluoride, which are separated by recrystallisation. In a modification of this process, potassium chloride is added to the hydrofluoric acid solution of the niobic and tantalic acids instead of potassium fluoride; the double potassium niobium fluoride, K₂NbF₇, remains dissolved and the tantalum salt is precipitated.

Ruff and Schiller determined the solubilities of the double fluorides of niobium and tantalum, K₂NbF₇ and K₂TaF₇, in varying quantities of hydrofluoric acid and potassium fluoride, and based a method of fractional separation on the results, which showed that the solubility of both fluorides diminishes with increasing concentration of potassium fluoride and decreasing concentration of hydrofluoric acid; the solubility increases rapidly with rising temperature, and is always greater for the niobium than for the tantalum salt. The tantalum may also be precipitated and removed partly or completely as potassium tantalum oxyfluoride, while the niobium remains in solution. Conversion of the separated fluorides into the corresponding oxides is effected by boiling with concentrated sulphuric acid until free from fluorine, and then hydrolysing the product by boiling with water. Alternatively, the hydrated acids are precipitated by the addition of ammonia to the solutions of the double fluorides. Niobium pentoxide, Nb₂O₅, or tantalum pentoxide, Ta₂O₅, is obtained on ignition of the precipitated hydrate.

A method of separation which avoids the preparation of the double fluorides consists in fusing the mixed niobic and tantalic acids with sodium carbonate and nitrate; the product is digested with warm water and a current of carbon dioxide is passed through the solution. It is claimed that only tantalic acid is precipitated. This process has, however, been the subject of adverse criticism. Partial separation of niobium from tantalum can be effected by warming the mixed, freshly precipitated, hydrated oxides with a mixture of hydrogen peroxide and hydrochloric acid; the niobium dissolves readily, while the tantalum dissolves only sparingly.

A more recent process, which avoids the difficulties associated with Marignac's method, is based on the solubility of niobium pentoxide and the comparative insolubility of tantalum pentoxide in a mixture of equal volumes of selenium oxychloride, SeOCl₂, and concentrated sulphuric acid. The tantalum pentoxide is left in the residue, and hydrolysis of the extract after dilution yields niobic acid.