I. Separation of the alcohols from other substances. - Before the alcohol in a mixture can be examined as to its character, or its quantity accurately determined, it must, as a rule, be obtained practically free from all other substances except water.

Simple distillation suffices for this where the admixed or dissolved ingredients are not volatile with steam. When volatile bodies are present, however, it is necessary either to render them incapable of distillation, or to remove them. Solutions containing ammonia or amines, for instance, may be made acid with sulphuric acid; volatile acids may be fixed by the addition of sodium or potassium hydroxide; free iodine may be converted into sodium iodide by treatment with sodium thiosulphate. Neutral organic volatile substances in general are best removed by a preliminary extraction with an immiscible solvent, or else by distillation, extraction of the distillate, and re-distillation.

For such articles as spirituous medicines, tinctures, essences, and perfumes, the method most generally applicable is that described by Thorpe and Holmes.1 By this process essential oils, camphor, ether, chloroform, ethyl acetate, and most other volatile constituents which would vitiate the subsequent examination of the distillate, are removed, either entirely or with sufficient completeness for the purpose. It is devised especially for the accurate determination of the proportion of alcohol; if this is already known or not required, some of the directions may obviously be modified.

Twenty-five c.c. of the sample, measured at 155°, are mixed with water in a separator to a bulk of 100 to 150 c.c, and common salt is added in sufficient quantity to saturate the liquid. The mixture is now shaken vigorously for five minutes with from 50 to 80 c.c. of light petroleum, boiling below 60°, and after standing review of the chief reactions which had been used for the detection of methyl alcohol previously to the year 1905 has been given by H. Scudder.2 Some of the main points are mentioned in the section which follows.

1 Trans. Chem. Soc, 1903, 83, 314.

1T0 for about half an hour, the lower layer is drawn off into another separator, extracted if necessary a second time with petroleum and then drawn off into a distillation flask. Meanwhile, the petroleum layers are washed successively with 25 c.c. of saturated common salt solution, and the washings added to the main bulk, which is neutralised if necessary, and then distilled and the distillate made up to 100 c.c.

In this description the article examined is assumed to be strongly spirituous, which is why it is diluted, and distilled to four times the volume taken. A weakly alcoholic liquid would not, of course, need to be diluted; 100 or 150 c.c. would be taken in the first instance.

Discussing the quantitative aspect of the process, the authors make the following comments.

On account of the high vapour pressure of ethyl alcohol and its avidity for water, the distillation of a strong spirituous liquid into its own volume of water so as to obtain an accurate determination of the amount of alcohol present is practically impossible with the apparatus and the methods of distillation commonly used. It is practicable to distil a strong spirituous liquid into twice its own volume with the requisite degree of accuracy, whilst distillation into four times the initial volume is quite easy, and the results are uniformly accurate. There is practically no error due to distillation, and as regards extent of dilution, it was found that the alcohol in 25 c.c. of strong spirit when diluted with water even to the extent of 500 c.c. and the solution saturated with salt, could be entirely recovered in the first 100 c.c. of the distillate.

The method, as described, is applicable to preparations containing chloroform, ether, benzaldehyde, and esters. In the greater number of other cases - for example, essences of lemon, juniper, peppermint, and santal oil preparations - a single extraction with petroleum ether is sufficient. For camphor preparations, 25 c.c. of normal sulphuric acid solution are used instead of common salt, and a single extraction with light petroleum is made, the acid being neutralised with caustic soda solution before the alcohol is distilled off.

Special devices are occasionally necessary to eliminate volatile foreign bodies from alcoholic liquids. Chloral, for instance, may be removed by heating the solution with caustic soda under a reflux condenser, the chloroform thus produced being afterwards extracted with light petroleum as usual. Coumarin, again, which distils to some extent with alcohol, is converted into the sodium salts of coumaric and coumarinic acids by the same treatment, and can thus be eliminated. Sometimes, too, extraction of a substance can advantageously be made with chloroform, carbon tetrachloride, ether or other solvent, the residue of which in its turn may have to be extracted with petroleum ether.

Another method of eliminating volatile oils and some similar substances is to dilute the sample until it contains about 20-25 per cent. of alcohol by volume, shake it well with powdered magnesium carbonate, and filter. The following example of a combination of this treatment with subsequent chemical removal of acetone is given by Leach and Lythgoe,1 for the case of a denatured alcohol containing light petroleum and acetone.

Take 25 c.c. of the sample and dilute it to 100 c.c. with water. Add about 5 grams of powdered magnesium carbonate, shake well, and filter. The filtrate will be free from petroleum, but will still contain the acetone and alcohols.

A measured portion of the filtrate, 55 c.c, is washed into a distilling flask and treated with 10 grams of powdered potassium bisulphite, and after standing an hour is distilled, 55 c.c. of distillate being collected. This is free from acetone, but contains some sulphurous acid, and in order to remove this it is distilled with sodium hydroxide, the final distillate being made up to 55 c.c. (for the determination of specific gravity and refraction).

II. Concentrating the alcohol.- For certain of the reactions described further on, the alcohol is required to be of about 90 per cent, strength, or more. Where the iodides are prepared, for example, as in the nitromethane test for methyl alcohol or in the Riche and Bardy process (p. 192), 10 c.c. of strong alcohol are used.

A ready means of obtaining this from a weak distillate, without using a fractionating column, is to take such a quantity of the liquid as will give 11 or 12 c.c. of the required strong alcohol, saturate it with common salt, and distil off about one-third. A liberal quantity of dry potassium carbonate is then added to the distillate, the mixture re-distilled, and the requisite quantity of distillate collected. The potassium carbonate forms with the water present a heavy layer under the separated alcohol, and the latter can be distilled off at high strength.

Suppose, for example, that a strong tincture or essence containing about 90 per cent. of alcohol has been purified by the Thorpe and Holmes method as described in the previous section, and it is now required to apply one of the above-mentioned tests to the distillate. Since 25 c.c. of the sample have been diluted to 100 c.c.

1 J. Amer. Chem. Soc, 1905, 27, No. 8.

the alcoholic strength of the distillate will be about 22.5 per cent. One-half of the 100 c.c. will yield more than sufficient spirit for one of the experiments in question. Hence 50 c.c. are taken, placed in a flask, saturated with salt, and 16 or 17 c.c. distilled off; the quantity thus obtained is redistilled from potassium carbonate and 10-11 c.c. collected. This will be of suitable strength for conversion into iodide.

For concentrating very dilute solutions - containing, say, a few tenths per cent. of alcohol - the liquid may be partly saturated with either calcium chloride or sodium chloride, one-sixth to one-third distilled off, and the operation repeated if necessary on the distillate. Thus R. F. Bacon found that 500 c.c. of a 0 2 per cent. solution, after addition of 200 grams of calcium chloride, gave on distillation 96.8 per cent. of its alcohol in the first 100 c.c. of distillate, and the whole in the first 150. A similar solution, about three-fourths saturated with sodium chloride, also gave up 96.8 per cent. in the first 100 c.c.; and on similarly salting the latter quantity and redistilling, the alcohol was obtained without further loss in the first 50 c.c. of distillate.

Of a still weaker solution, containing only 0 1 per cent. of alcohol, 1 litre was taken, and distilled after being about three-fourths saturated with sodium chloride. The first 150 c.c. of distillate were similarly salted and redistilled, with the result that 96 per cent. of the alcohol originally present was recovered in the first 25 c.c. of the second distillate.1

By one or more of the operations indicated above, ethyl and methyl alcohols are separated from the other constituents of mixtures in a condition of sufficient purity and strength for further examination. The final distillate obtained may of course contain either or both of these two alcohols.