where t′ indicates the temperature of water corresponding with a vapour tension p, and where c is the ratio of 273 + t′ to 273 + t. The magnitude of c is evidently expressed with great accuracy by the straight line c = 1·1703 + 0·0011t. In exactly the same way we find the ratio for liquid bromine and water to be c1 = 1·1585 + 0·00057t. The intersection of these straight lines in fact corresponds with -7°·06, which again confirms the melting point given above for bromine. In this manner it is possible with the existing store of data to accurately establish and verify the melting point of substances. Ramsay and Young established the thermal constants of iodine by exactly the same method.
[60] The observations made by Paterno and Nasini (by Raoult's method, Chapter I. Note [49]) on the temperature of the formation of ice ( -1°·115, with 1·391 gram of bromine in 100 grams of water) in an aqueous solution of bromine, showed that bromine is contained in solutions as the molecule Br2. Similar experiments conducted on iodine (Kloboukoff 1889 and Beckmann 1890) show that in solution the molecule is I2.
B. Roozeboom investigated the hydrate of bromine as completely as the hydrate of chlorine (Notes [9], [10]). The temperature of the complete decomposition of the hydrate is +6°·2; the density of Br2,10H2O = 1·49.
[61] In general, 2HI + O = I2 + H2O, if the oxygen proceed from a substance from which it is easily evolved. For this reason compounds corresponding with the higher stages of oxidation or chlorination frequently give a lower stage when treated with hydriodic acid. Ferric oxide, Fe2O3, is a higher oxide, and ferrous oxide, FeO, a lower oxide; the former corresponds with FeX3, and the latter with FeX2, and this passage from the higher to the lower takes place under the action of hydriodic acid. Thus hydrogen peroxide and ozone (Chapter [IV].) are able to liberate iodine from hydriodic acid. Compounds of copper oxide, CuO or CuX2, give compounds of the suboxide Cu2O, or CuX. Even sulphuric acid, which corresponds to the higher stage SO3, is able to act thus, forming the lower oxide SO2. The liberation of iodine from hydriodic acid proceeds with still greater ease under the action of substances capable of disengaging oxygen. In practice, many methods are employed for liberating iodine from acid liquids containing, for example, sulphuric acid and hydriodic acid. The higher oxides of nitrogen are most commonly used; they then pass into nitric oxide. Iodine may even be disengaged from hydriodic acid by the action of iodic acid, &c. But there is a limit in these reactions of the oxidation of hydriodic acid because, under certain conditions, especially in dilute solutions, the iodine set free is itself able to act as an oxidising agent—that is, it exhibits the character of chlorine, and of the halogens in general, to which we shall again have occasion to refer. In Chili, where a large quantity of iodine is extracted in the manufacture of Chili nitre, which contains NaIO3, it is mixed with the acid and normal sulphites of sodium in solution; the iodine is then precipitated according to the equation 2NaIO3 + 3Na2SO3 + 2NaHSO3 = 5Na2SO4 +I2 + H2O. The iodine thus obtained is purified by sublimation.
[62] For the final purification of iodine, Stas dissolved it in a strong solution of potassium iodide, and precipitated it by the addition of water (see Note [58]).
[63] The solubility of iodine in solutions containing iodides, and compounds of iodine in general, may serve, on the one hand, as an indication that solution is due to a similarity between the solvent and dissolved substance, and, on the other hand, as an indirect proof of that view as to solutions which was cited in Chapter [I]., because in many instances unstable highly iodised compounds, resembling crystallo-hydrates, have been obtained from such solutions. Thus iodide of tetramethylammonium, N(CH3)4I, combines with I2, and I4. Even a solution of iodine in a saturated solution of potassium iodide presents indications of the formation of a definite compound KI3. Thus, an alcoholic solution of KI3 does not give up iodine to carbon bisulphide, although this solvent takes up iodine from an alcoholic solution of iodine itself (Girault, Jörgensen, and others). The instability of these compounds resembles the instability of many crystallo-hydrates, for instance of HCl,2H2O.
[64] The equality of the atomic volumes of the halogens themselves is all the more remarkable because in all the halogen compounds the volume augments with the substitution of fluorine by chlorine, bromine, and iodine. Thus, for example, the volume of sodium fluoride (obtained by dividing the weight expressed by its formula by its specific gravity) is about 15, of sodium chloride 27, of sodium bromide 32, and of sodium iodide 41. The volume of silicon chloroform, SiHCl3, is 82, and those of the corresponding bromine and iodine compounds are 108 and 122 respectively. The same difference also exists in solutions; for example, NaCl + 200H2O has a sp. gr. (at 15°/4°) of 1·0106, consequently the volume of the solution 3,658·5/1·0106 = 3,620, hence the volume of sodium chloride in solution = 3,620–3,603 (this is the volume of 200 H2O) = 17, and in similar solutions, NaBr = 26 and NaI = 35.
[65] But the density (and also molecular volume, Note [64]) of a bromine compound is always greater than that of a chlorine compound, whilst that of an iodine compound is still greater. The order is the same in many other respects. For example, an iodine compound has a higher boiling point than a bromine compound, &c.
[66] A. L. Potilitzin showed that in heating various metallic chlorides in a closed tube, with an equivalent quantity of bromine, a distribution of the metal between the halogens always occurs, and that the amounts of chlorine replaced by the bromine in the ultimate product are proportional to the atomic weights of the metals taken and inversely proportional to their equivalence. Thus, if NaCl + Br be taken, then out of 100 parts of chlorine, 5·54 are replaced by the bromine, whilst with AgCl + Br 27·28 parts are replaced. These figures are in the ratio 1 : 4·9, and the atomic weights Na : Ag = 1 : 4·7. In general terms, if a chloride MCln be taken, it gives with nBr a percentage substitution = 4M/n2 where M is the atomic weight of the metal. This law was deduced from observations on the chlorides of Li, K, Na, Ag (n = 1), Ca, Sr, Ba, Co, Ni, Hg, Pb (n = 2), Bi (n = 3), Sn (n = 4), and Fe2 (n = 6).
In these determinations of Potilitzin we see not only a brilliant confirmation of Berthollet's doctrine, but also the first effort to directly determine the affinities of elements by means of displacement. The chief object of these researches consisted in proving whether a displacement occurs in those cases where heat is absorbed, and in this instance it should be absorbed, because the formation of all metallic bromides is attended with the evolution of less heat than that of the chlorides, as is seen by the figures given in Note [55].