flexible piece of apparatus. The side-arm of the separator may be made with two rubber connections,— one above and one below the tube leading to the potassium carbonate tube. The long return tube to the flask may be constructed with a rubber joint very near the carbonate tube and one near the flask.

3. Other Methods of Preparation

Ethyl oxalate has been prepared in poor yields by the following methods: by distilling a mixture of anhydrous oxalic acid and absolute alcohol;[1] by heating a mixture of anhydrous oxalic acid and 97 per cent alcohol under a reflux condenser and fractionating the resulting mixture;[2] by distilling a mixture of anhydrous oxalic acid and absolute alcohol, the vapor of absolute alcohol being passed simultaneously into the mixture;[3] by allowing a saturated solution of oxalic acid in alcohol to stand for a long time at 40-50'0.[4]

A good yield has been obtained by Anschutz[5] by a method involving saturation of a mixture of crystallized oxalic acid and alcohol with hydrogen chloride, removal of the alcohol and water by distillation under reduced pressure, and repetition of the treatment with the alcohol and hydrogen chloride, the process being carried out several times.

[1] Jahresb. 1861, 598.

[2] J prakt. Chem. (2), 34, 500 (1886).

[3] Monatsh. 17, 614 (1896).

[4] Ann. 65, 350 (1848).

[5] Ber. 16, 2414 (1883),

VII

ETHYL PHENYLACETATE

C6H5CH2CN + C2H5OH + H2SO4 + H2O—> C6H5CH2CO2C2H5 + NH4HSO4

Prepared by ROGER ADAMS and A. F. THAL. Checked by OLIVER KAMM.

1. Procedure

IN a 3-l. round-bottom flask, fitted with an efficient reflux condenser, are mixed 750 g. of 95 per cent alcohol, 750 g. of concentrated sulfuric acid and 450 g. of benzyl cyanide. The mixture, which soon separates into two layers, is heated to boiling over a low flame, for six to seven hours, cooled and poured into 2 l. of water, and the upper layer is separated. This is washed with a little 10 per cent sodium carbonate solution to remove small amounts of phenylacetic acid which may have been formed, and then distilled in vacuo. A small amount of water goes over first and then a pure product boiling 132-138'0/32 mm. (120-125'0/17-18 mm.). The yield varies in general between 525 and 550 g. (83-87 per cent of the theoretical amount).

2. Notes

The benzyl cyanide can be most conveniently prepared according to the directions in preparation III (p. 9); the product which boils over a 5'0 range should be used.

In washing the layer of ethyl phenylacetate with sodium carbonate it is sometimes advisable to add a certain amount of sodium chloride so that the ester will separate more readily.

The product obtained is water-clear and practically colorless. Although the product is collected over a 5'0 range, most of the liquid is found to boil over a 1'0 range, if distilled slowly without superheating.

The boiling point of ethyl phenylacetate is near that of benzyl cyanide. However, a Kjeldahl analysis of the product shows that only a trace of nitrogen compounds is present.

3. Other Methods of Preparation

Ethyl phenylacetate may be prepared by the treatment of benzyl cyanide with alcohol and hydrochloric acid gas.[1] It is much more convenient in the laboratory, however, to use sulfuric acid in place of hydrochloric acid; in fact, the yields obtained are better than those recorded in the literature. This ester may also be made by the esterification of phenylacetic acid with hydrochloric acid and alcohol;[2] or with alcohol and sulfuric acid;[3] the following less important methods of preparation may be mentioned; the action of benzyl magnesium chloride upon ethyl chlorocarbonate,[4] and the action of copper on a mixture of bromobenzene and ethyl chloroacetate at 180'0.[5]

[1] Ber. 20, 592 (1887); Ann. 296, 361 (1897)

[2] Ber. 2, 208 (1869).

[3] Ann. 296, 2, footnote (1897); Compt. rend. 152, 1855 (1911).

[4] Ber. 36, 3088 (1903).

[5] Ber. 2, 738 (1869).

VIII

GLYCEROL a, g-DICHLOROHYDRIN

C3H5(OH)3 + 2HCl—> CH2ClCHOHCH2Cl + 2H2O

Prepared by J. B. CONANT and O. R. QUAYLE. Checked by O. KAMM and A. O. MATTHEWS.

1. Procedure

ONE kilo of 90 per cent glycerol (sp. gr. 1.243) and 20 g. of acetic acid are placed in a weighed 2-l. flask which is immersed in an oil bath heated to 100-110'0. The flask is fitted with a two-hole stopper, which carries a long tube reaching to the bottom of the flask and a short exit tube. The former is connected to a hydrogen chloride generator, the latter to a catch-bottle and some system for absorbing any excess of hydrogen chloride. A stream of dry hydrogen chloride is passed into the mixture. The absorption of gas is very rapid at the start, but gradually falls off towards the end of the reaction; the stream of hydrogen chloride should be regulated accordingly. The flask is removed from time to time and weighed; when the absorption of gas practically ceases, the increase in weight will be about 875 g. (25 per cent more than the theoretical amount).

The product is now cooled, placed in a 4-l. beaker, and treated with solid sodium carbonate until just alkaline to litmus. Water is added from time to time, to facilitate the reaction with the sodium carbonate and to prevent the separation of salt; about 500 cc. are required. The mixture is transferred to a separatory funnel and the aqueous layer separated. The crude dichlorohydrin, which weighs 1250 g., is distilled in vacuo. The first fraction boiling below 68'0/14 mm. weighs 225 g., and consists of water and some dichlorohydrin; the dichlorohydrin is collected between 68-75'0/14 mm., and weighs about 775 g. The water is separated from the first fraction, which is then redistilled and yields 100 g. of dichlorohydrin. A still further amount of material (40-45 g.) may be obtained by extracting with benzene, the aqueous layer obtained in the neutralization process. This is, however, hardly profitable. The neutralization and distillation will require about four hours.

The 875 g. of dichlorohydrin thus obtained boils over a 7'0 range; this is 70 per cent of the theoretical amount. Redistillation yields 700-720 g. boiling 70-73'0/14 mm. (57 per cent of the theoretical amount).

2. Notes

The most convenient hydrogen chloride generator is that described by Sweeney.[1] Concentrated hydrochloric acid is introduced into concentrated sulfuric acid, by means of a dropping funnel and a capillary tube leading to the bottom of the sulfuric acid container. It is convenient to use a 3-l. bottle for this container and a 1-l. funnel to contain the hydrochloric acid. The gas is dried by passing through a wash-bottle containing concentrated sulfuric acid. An empty catch-flask should be connected between the generator and the absorption flask in case any glycerol tends to suck back at the start of the reaction. About 6 kg. of concentrated hydrochloric acid and 10 kg. of concentrated sulfuric acid are required in one run. The generating flask will have to be recharged every six hours; it should be half filled with sulfuric acid. Aside from this, the apparatus needs no attention. The oil bath can be conveniently heated on an electric hot plate.

The dichlorohydrin boiling