In preparing anhydrous alcohol, the usual procedure is to digest 95 per cent. spirit with quicklime for some days, and then to distil off the alcohol slowly from a water-bath and treat the product in the same way. The first and last portions of the distillate are rejected. A sufficient excess of lime must be used, otherwise too large a proportion of calcium hydroxide is formed, and towards the end of the distillation this yields up water and thus dilutes the alcohol again. About one-fourth or one-fifth the weight of the alcohol is a suitable quantity of lime for the distillation.

If strong spirit is not available to start with, it may readily be obtained, of about 94 or 95 per cent. strength, by distilling weak alcohol from potassium carbonate. The strength may be raised to about 97 per cent. by distilling the product over fused calcium chloride, and the distillate may then be treated with quicklime to obtain the absolute alcohol.

Barium oxide may also be used as a dehydrating agent, and powdered calcium carbide has been employed.2 Simple percolation of strong alcohol through a column of quicklime, after digestion with the latter for several days, yields a nearly anhydrous product - containing, however, a small quantity of lime, which may be removed by distillation. Metallic sodium and metallic calcium have been used for removing the last traces of water, but quicklime appears to be, on the whole, the best dehydrating agent of all. Its action may be accelerated by heating the mixture under a reflux condenser, and by using a larger proportion of the lime - up to one-half the weight of the alcohol - in the first treatment.

1 Compt. rend., 1910, 150, 823. 2 Yvon, Chem. News, 1898, 52.

The following details of an actual experiment are given by R. W. Merriman.1

(a) Ten litres of 95 per cent. alcohol were boiled in a copper drum for twenty-four hours with 4 kilos. of quicklime made from marble. On distilling the product, spirit containing about 0 5 per cent. of water was obtained. This formed the stock, which was further dehydrated as required in batches of 700 grams.

(6) Seven hundred grams of the 99.5 per cent. alcohol were boiled on the water-bath, under a reflux condenser, with 150 grams of quicklime, taking the usual precautions for preventing the absorption of atmospheric moisture. After six hours, the product was distilled from the water-bath, a double spray trap being fixed on to the flask. The first 50 grams were rejected; the distillate obtained by boiling almost to dryness (about 600 grams) had a density of 080630 at 0°/4°, and thus contained 0.006 per cent, of water.

(c) The alcohol of sp. gr. 0.80630 was again treated with lime, and the density of the main fraction was 0.80628 at 0°/4°. Another treatment with lime, and also with metallic calcium, left the density unchanged.

The density of the product, it may be remarked, is slightly higher than that of Mendeleeff's alcohol; but the experiment shows, at all events, that a very nearly anhydrous alcohol is given by the treatment described.

Young's method of dehydrating alcohol by distillation with benzene has already been referred to (p. 157).

Anhydrous potassium fluoride is a good preliminary dehydrant for alcohol. It acts rapidly on account of its solubility, and is somewhat more effective than potassium carbonate, as the saturated solution has a lower vapour pressure. It absorbs a large proportion of water (62 per cent.), forming the hydrate KF,2H1O, the vapour pressure of which is 3 5 mm., and the limit of dehydration of ethyl alcohol is reached at 97.5 per cent., when the vapour pressure corresponds with that of the above hydrate.2

According to L. W. Winkler,3 alcohol can be completely dehydrated by distilling it over filings of metallic calcium. The latter may contain some nitride, which will yield ammonia by the action of water in the alcohol; but this ammonia can be eliminated afterwards. The filings are sifted in a coarse sieve to free them from any powdered nitride, and traces of oil and more nitride are removed by shaking the filings vigorously with dry carbon tetrachloride in a glass-stoppered flask, transferring them to a loosely-stoppered funnel, washing them again with the same solvent, and then drying in carbon dioxide. The alcohol is distilled with the purified metal in the proportion of 20 grams to the litre. Ammonia is removed from the distillate by dissolving in the latter a little alizarin (a few centigrams per litre) and then adding a solution of about 0 5 gram of tartaric acid in 10 c.c. of the distilled alcohol until the reddish-blue colour has turned to a pure yellow. After the addition of a few more drops of the acid solution when this point has been reached, the alcohol is re-distilled, care being taken to exclude atmospheric moisture during the operation.

1 Trans. Ghem. Soc, 1913, 103, 629.

2 G. B. Frankforter and F. C. Frary, J. Phys. Chem., 1913, 17, 402.

3 Zeitsch. angew. Chem., 1916, 29, 18.

Perkin and Pratt have found that when metallic calcium is allowed to remain in contact with ethyl (or methyl) alcohol for a period of 30 to 60 minutes, a reaction occurs between the metal and the alcohol.1 Once started, the reaction, which may become very vigorous, continues to completion, with formation of calcium ethoxide (or methoxide) according to the equation: -

2C2H5+OH + Ca = (C2H5O)2Ca + H1.

If the mixture is heated to the boiling point of the alcohol, the reaction proceeds much more rapidly.

L. W. Andrews2 found that absolute alcohol, dehydrated by means of lime freshly prepared from marble, and freed from aldehyde, had the same specific gravity, the same refractive index, and the same critical temperature of solution, as alcohol had which was dehydrated with magnesium amalgam 3 or metallic calcium. The specific gravity found was, at 25°/4°, 078510 ± 0.00001.

Winkler, in 1905, obtained the value 078509 at 25°/4° for the specific gravity of absolute alcohol dehydrated by means of metallic calcium. A. Kailan4 later found a somewhat higher value, namely, 0.78513 at 25°/4°, using the same method of dehydration. Mende-leeff's absolute alcohol was dehydrated by means of quicklime: its specific gravity referred to the above temperatures (25°/4°) was 078522. Referred to the temperature 15.6o/15.6° the value is 0.79359, and as already stated, this is regarded as probably the most accurate result.

According to a German patent,5 alcohol of 90 per cent. strength or more can be dehydrated completely in a short time by means of anhydrous sodium sulphide. This compound is indifferent towards ethyl alcohol, and can be entirely recovered by heating after use.

1 Proc. Chem. Soc., 1907, 23, 304. 2 J. Amer. Chem. Soc., 1908, 30, 353. 3 Ibid., 1904, 26, 1158. 4 Ber., 1911, 44, 2881.

5 D.R.-P. 256591 (1909).

Commercial ' absolute ' alcohol usually contains water in quantity ranging from about 0 5 to 1.5 per cent. The "absolute alcohol ' of the British Pharmacopoeia is defined as " ethyl hydroxide, C2H5.OH, with not more than 1 per cent. by weight of water. Specific gravity (at 15.5°/15.5°) from 0.794 (equivalent to 99.95 per cent. of ethyl hydroxide by volume and by weight) to 0.7969 (equivalent to 99.4 per cent. of ethyl hydroxide by volume or 99 per cent. by weight)." More accurate values for the specific gravity would be 0.7938 and 0.7967 respectively. The test prescribed to exclude excess of water is as follows. " Anhydrous copper sulphate shaken occasionally during two or three hours with about fifty times its weight of absolute alcohol does not assume a decidedly blue colour."

Other tests for water in alcohol are given in Chap. VI, and also methods for the examination of alcohol as regards the presence of impurities.

According to the French Pharmacopoeia, commercial "absolute" alcohol should have a specific gravity not higher than 0.79683 at 15°/15°.

For the benefit of microscopists and other private workers who have no chemical balance or hydrometer, H. Garnett mentions that pure oil of cedarwood (Juniperus virginiana) may be employed for testing the strength of absolute alcohol.1 (The so-called " thickened " oil used for immersion lenses will not serve.) With absolute alcohol, the oil mixes clear in all proportions at the ordinary temperature. An alcohol of 98.2 per cent. strength by volume mixes without turbidity with cedarwood oil at 15.5°, but turbidity is produced if the temperature is lowered even 1°. An alcohol of this strength will suffice for microscopical work.

Calcium carbide can be used to increase the strength of an alcohol. Thus alcohol of 98.9 per cent. strength (by volume) was digested with a small quantity of powdered carbide for three days in a warm place. Some acetylene was evolved, and the strength was raised to 99.94 per cent. Any disagreeable odour left may be removed by shaking the alcohol with a minute trace of potassium permanganate.