Such acid should give no precipitate, either with a solution of barium chloride (a precipitate shows the presence of sulphuric acid) or with a solution of silver nitrate (a precipitate shows the presence of hydrochloric acid), nor should it, after being diluted with water, give a coloration with starch containing potassium iodide (a coloration shows the admixture of other oxides of nitrogen). The oxides of nitrogen may be most easily removed from impure nitric acid by heating for a certain time with a small quantity of pure charcoal. By the action of nitric acid on the charcoal carbonic anhydride is evolved, which carries off the lower oxides of nitrogen. On redistilling, pure acid is obtained. The oxides of nitrogen occurring in solution may also be removed by passing air through the nitric acid.
Dalton, Smith, Bineau, and others considered that the hydrate of constant boiling point (see Chapter I., Note 107) for nitric acid was the compound 2HNO3,3H2O, but Roscoe showed that its composition changes with a variation of the pressure and temperature under which the distillation proceeds. Thus, at a pressure of 1 atmosphere the solution of constant boiling point contains 68·6 p.c., and at one-tenth of an atmosphere 66·8 p.c. Judging from what has been said concerning solutions of hydrochloric acid, and from the variation of specific gravity, I think that the comparatively large decrease in the tensions of the vapours depends on the formation of a hydrate, NHO3,2H2O (= 63·6 p.c.). Such a hydrate may be expressed by N(HO)5, that is, as NH4(HO), in which all the equivalents of hydrogen are replaced by hydroxyl. The constant boiling point will then be the temperature of the decomposition of this hydrate. The variation of the specific gravity at 15° from water (p = 0) to the hydrate NHO3,5H2O (41·2 p.c. HNO3) is expressed by s = 9992 + 57·4p + 0·16p2, if water = 10,000 at 4°. For example, when p = 30 p.c., s = 11,860.