[38] Since the close of the eighties in England, where the preparation of sodium is at present carried out on a large commercial scale (from 1860 to 1870 it was only manufactured in a few works in France), it has been the practice to add to Deville's mixture iron, or iron oxide which with the charcoal gives metallic and carburetted iron, which still further facilitates the decomposition. At present a kilogram of sodium may be purchased for about the same sum (2/-) as a gram cost thirty years ago. Castner, in England, greatly improved the manufacture of sodium in large quantities, and so cheapened it as a reducing agent in the preparation of metallic aluminium. He heated a mixture of 44 parts of NaHO, and 7 parts of carbide of iron in large iron retorts at 1,000° and obtained about 6½ parts of metallic sodium. The reaction proceeds more easily than with carbon or iron alone, and the decomposition of the NaHO proceeds according to the equation: 3NaHO + C = Na2CO3 + 3H + Na. Subsequently, in 1891, aluminium was prepared by electrolysis (see Chapter XVII.), and metallic sodium found two new uses; (1) for the manufacture of peroxide of sodium (see later on) which is used in bleaching works, and (2) in the manufacture of potassium and sodium cyanide from yellow prussiate (Chapter XIII., Note [12]).

[38 bis] This is also shown by the fall in the temperature of solidification of tin produced by the addition of sodium (and also Al and Zn). Heycock and Neville (1889).

[39] By dissolving sodium amalgams in water and acids, and deducting the heat of solution of the sodium, Berthelot found that for each atom of the sodium in amalgams containing a larger amount of mercury than NaHg5, the amount of heat evolved increases, after which the heat of formation falls, and the heat evolved decreases. In the formation of NaHg5 about 18,500 calories are evolved; when NaHg3 is formed, about 14,000; and for NaHg about 10,000 calories. Kraft regarded the definite crystalline amalgam as having the composition of NaHg6, but at the present time, in accordance with Grimaldi's results, it is thought to be NaHg5. A similar amalgam is very easily obtained if a 3 p.c. amalgam be left several days in a solution of sodium hydroxide until a crystalline mass is formed, from which the mercury may be removed by strongly pressing in chamois leather. This amalgam with a solution of potassium hydroxide forms a potassium amalgam, KHg10. It may be mentioned here that the latent heat of fusion (of atomic quantities) of Hg = 360 (Personne), Na = 730 (Joannis), and K = 610 calories (Joannis).

[40] Alloys are so similar to solutions (exhibiting such complete parallelism in properties) that they are included in the same class of so-called indefinite compounds. But in alloys, as substances passing from the liquid to the solid state, it is easier to discover the formation of definite chemical compounds. Besides the alloys of Na with Hg, those with tin (Bailey 1892 found Na2Sn), lead (NaPb), bismuth (Na3Bi), &c. (Joannis 1892 and others) have been investigated.

[41] Potassium forms a similar compound, but lithium, under the same circumstances, does not.

[42] The tension of dissociation of hydrogen p, in millimetres of mercury, is:—

t =330°350°400° 430°
for Na2Hp = 28 57447 910
for K2H 45 725481100

[43] In general, during the formation of alloys the volumes change very slightly, and therefore from the volume of Na2H some idea may be formed of the volume of hydrogen in a solid or liquid state. Even Archimedes concluded that there was gold in an alloy of copper and gold by reason of its volume and density. From the fact that the density of Na2H is equal to 0·959, it may be seen that the volume of 47 grams (the gram molecule) of this compound = 49·0 c.c. The volume of 46 grams of sodium contained in the Na2H (the density under the same conditions being 0·97) is equal to 47·4 c.c. Therefore the volume of 1 gram of hydrogen in Na2H is equal to 1·6 c.c., and consequently the density of metallic hydrogen, or the weight of 1 c.c., approaches 0·6 gram. This density is also proper to the hydrogen alloyed with potassium and palladium. Judging from the scanty information which is at present available, liquid hydrogen near its absolute boiling point (Chapter [II].) has a much lower density.

[43 bis] We may remark that at low temperatures Na absorbs NH3 and forms (NH3Na)2 (see Chapter VI., Note [14]); this substance absorbs CO and gives (NaCO)n (Chapter IX., Note [31]), although by itself Na does not combine directly with CO (but K does).

[44] H. A. Schmidt remarked that perfectly dry hydrogen chloride is decomposed with great difficulty by sodium, although the decomposition proceeds easily with potassium and with sodium in moist hydrogen chloride. Wanklyn also remarked that sodium burns with great difficulty in dry chlorine. Probably these facts are related to other phenomena observed by Dixon, who found that perfectly dry carbonic oxide does not explode with oxygen on passing an electric spark.