A cetonitrile.

A solution of 1.08 grams of the l-bromo-acid, [a] −8.8°, in 15 c.c. of acetonitrile was saturated with dry ammonia at 9°, whereby the volume increased by 1.2 c.c. The solution at first was clear, but a precipitate consisting of the ammonium salts of the amino-acid, the imino-acid, and cinnamic acid soon began to form. After twentyfour days, the precipitate was collected, and the amino-acid obtained as usual, the yield being 0-2 gram, [a+9°. The imino-acid was accidentally lost.

A second experiment with a less active bromo-acid ([a]5+5·8°) in more concentrated solution (3·9 grams in 17 c.c. of solvent), kept for twenty-one days, gave 0.35 gram of (recrystallised) phenylalanine, [a] -8°, and 0.57 gram of imino-acid (compare the section on the imino-acid). It appears therefore that, as would be anticipated, the more concentrated the solution of bromo-acid the greater is the yield of imino-acid.

The Alcohols.

As the reaction in these solvents is very slow, the solutions were kept for 280-300 days before separating the products. In each case, 0·7-0·8 gram of active bromo-acid was used, and the yield of amino-acid (except with benzyl alcohol) varied from 0.15-0.25 gram. With methyl alcohol, the solutions remained clear throughout, showing that the amino-acid and imino-acid (or their ammonium salts) are soluble in this solvent. With the other alcohols, except benzyl alcohol, a small amount of precipitate formed after a time. The absence of a precipitate in the benzyl alcohol solution is probably due to the slow rate of reaction in this solvent, as only 0·06 gram of amino-acid and 0·02 gram of iminoacid were obtained, although the mixture was kept for 309 days. The relative activities of the amino-acids from the different solvents are given in the table. The properties of the partly active iminoacids obtained from these solvents are briefly dealt with in the next section.

a-Imino-BB-diphenyldipropionic Acid.

As this acid has not been previously obtained, it is desirable to describe its properties in some detail. Earlier experiments having shown that a good yield of the imino-acid was obtained from acetonitrile, 84 grams of the r-bromo-acid were dissolved in 20 c.c. of

this solvent, the solution being saturated with dry ammonia and kept for twenty-eight days. The precipitate was then collected, washed with ether, and the inactive imino-acid separated as already described (p. 158). After crystallisation from a mixture of acetone and water, in which it is very sparingly soluble, the compound forms stellar aggregates of elongated leaflets or needles melting and decomposing at 262-264°, with previous discoloration and shrinkage at 235°. Before analysis, it was dried for some hours at 100°:

0.2588 gave 0-6545 CO2 and 0.1447 H2O. C=690; H=6.27. 0.5561 211 c.c. N2 (moist) at 18° and 771 mm. N=4.5. C18H1904N requires C=690; H=6·1; N=4.5 per cent.

a-Imino-BB-diphenyldipropionic acid is very sparingly soluble in water, practically insoluble in dilute or concentrated hydrochloric acid, but readily soluble in dilute alkali hydroxide solution. This shows that, as would be anticipated, it is an extremely weak base, whereas its acidic character is fairly pronounced. An attempt to determine its equivalent by evaporating with hydrochloric acid gave an unsatisfactory result, as the hygroscopic hydrochloride decomposed on drying at 100°. Further, no definite results for the equivalent were obtained by titration, using phenolphthalein as indicator, or by evaporating with ammonia, in an attempt to prepare a pure ammonium salt. Evidence was obtained in some of the experiments with acetonitrile as to the dibasic character of the acid, since the solid imino-acid compound which separates from this solvent during the reaction, on treatment with water in the cold, gives off ammonia, forming an ammonium salt (the hydrogen ammonium salt) soluble in water.

The amount of active imino-acid obtained from most of the solvents was very small, on the average 0.05 gram (from trimethylcarbinol 01 gram), but the rotation measurements were quite definite. The various specimens decomposed on heating to

245-251°.

Summary.

(1) The amino-acid obtained by the action of ammonia on optically active a-bromo-6-phenylpropionic acid in different solvents has in all cases the same sign, opposite to that of the original bromoacid.

(2) The amino-acid from liquid ammonia and acetonitrile is more racemised than that from other solvents.

(3) The presence of ammonium bromide has little effect on the

activity of the product from aqueous ammonia; it greatly diminishes the racemisation when the reaction is carried out in liquid ammonia.

(4) a-Imino-BB-diphenyldipropionic acid was isolated from the products of reaction in all solvents used, and its properties were investigated; the sign was in all cases the same as that of the original bromo-acid.

(5) The mechanism of displacement racemisation is discussed.

In conclusion, we desire to thank the Government Grant Committee of the Royal Society for a grant which has defrayed the expenses of this investigation.

CHEMISTRY DEPARTMENT,

BIRKBECK COLLEGE,

LONDON, 'E.C.

[Received, February 4th, 1918.]

XIX.-The Action of Aniline on Carbon
Tetrachloride.

BY ERNST JOHANNES HARTUNG.

THE action of aniline on carbon tetrachloride was studied long ago by Hofmann (Proc. Roy. Soc., 1858, 9, 284). He found that no action was apparent at the ordinary temperature, but that, on heating 3 volumes of aniline with 1 volume of carbon tetrachloride in a sealed tube for thirty hours at 170-180°, a soft, black mass was obtained, from which he prepared pararosaniline and new base, diphenylaminobenzamidine (carbotriphenyltriamine), C19H17N3. He also obtained the hydrochloride of this substance, but did not pursue his researches further. Weith (Ber., 1877, 10, 358) many years afterwards attacked the problem of the structure of the base. He prepared it by Hofmann's method, and succeeded in showing (Ber., 1879, 12, 101) that the constitution could be represented as

Later syntheses of diphenylaminobenzamidine (Michler and Walder, Ber., 1881, 14, 2174) confirmed the constitution thus established. The base is a product of the condensation of aniline with carbon tetrachloride, the latter supplying the central carbon atom. It is interesting in that the chlorine atoms of the tetrachloride have been displaced in three different ways: two chlorine

VOL. CXIII.

H

atoms have reacted with the two aminic hydrogen atoms in one aniline molecule, another has reacted with one aminic hydrogen atom in a second aniline molecule, and the fourth halogen atom has reacted with one of the ring hydrogen atoms in the paraposition.

The present paper is the outcome of some experiments on the reaction in the presence of a catalyst. During the course of certain other work, it was noticed that a mixture of pure aniline and pure carbon tetrachloride, colourless at first, rapidly darkened at the laboratory temperature when in contact with a copper-mercury couple, ultimately becoming almost solid. As the reaction appeared to be far more rapid than that studied by Hofmann, it was of interest to discover whether the same products were formed. With this end in view, mixtures of aniline and carbon tetrachloride were allowed to react under varying conditions, and the main products were investigated.

EXPERIMENTAL.

Mixture 1.-A mixture of 3 molecular proportions of pure aniline and 1 molecular proportion of pure carbon tetrachloride was placed in contact with a copper-mercury couple and allowed to remain in a closed vessel until solid; this required three days at the ordinary temperature. Repeated extraction of the resulting black, oily substance with hot water showed that much of the original mixture was still unchanged. A tenacious black tar remained which solidified, on cooling, to a brittle, shining solid.

The brownish-red filtrate was boiled until all odour of aniline had disappeared, decolorised with animal charcoal, and set aside to crystallise. The crystals were purified by repeated crystallisation from water, and were then obtained quite clean and white, melting at 279° (291° corr.). This melting point agrees with that of the diphenylaminobenzamidine hydrochloride obtained by Weith. The presence of ionisable chlorine was shown by precipitation with silver nitrate; the base itself was prepared by the action of dilute sodium hydroxide solution on the salt and repeated crystallisation of the product from alcohol. It formed soft, creamy-white needles melting at 1970 (203° corr.). (Found: C=79-10; H=6.16. C19H17Ng requires C-79-40; H=5.97 per cent.) The substance was thus undoubtedly diphenylaminobenzamidine.

The black solid separated earlier was suspended in water and distilled in a current of steam. The distillate had a strong odour of phenylcarbylamine and deposited orange, crystalline scales.

These, after separation and purification, were shown to be azobenzene by their melting point, reactions, and quantitative analysis. A few drops of phenylcarbylamine were separated from the filtrate by extraction with ether.

Mixture 2.-This mixture was the same in composition as the previous one, and was heated at 80-90° with the copper-mercury couple under a reflux condenser for fifty minutes, when the mass had the appearance of moist black sand. Repeated extraction with hot water left only a small amount of dark, sooty material behind. The brown filtrate contained some aniline and azobenzene, and was boiled until these were removed. On crystallisation, much diphenylaminobenzamidine hydrochloride was again obtained. The red mother liquor, containing undoubtedly some pararosaniline hydrochloride, was evaporated, but the red colour was gradually destroyed, and only a further amount of the salt previously obtained and some aniline hydrochloride were found.

It was noticed that the diphenylaminobenzamidine salt obtained from this mixture had a persistent pale yellow colour, in spite of attempts to remove it by means of animal charcoal. Further, the corresponding salt from Mixture 1, which was pure white, became yellow on firm rubbing against the side of the porcelain dish with a glass rod, but ultimately became white again after some days. These changes could be induced repeatedly, so that apparently two modifications of the salt exist. On determining the melting point, the substance turns yellow at about 230°, and this colour increases until melting at 279° (291° corr.) takes place to a red liquid with partial decomposition.

Mixture 3.-This was a mixture of about equal parts of aniline and carbon tetrachloride, part of which was the middle fraction in the separation of the two liquids by fractional distillation. It was allowed to remain for several months without heating and in the absence of the catalyst. At intervals of a few weeks, the crystals which had formed in the dark brown liquid were collected and the filtrate was returned to the bottle. In each case, the substance obtained was treated as follows: The felted, hair-like mass was pressed in absorbent paper, extracted with water, and the extract allowed to crystallise. When crystals appeared, they were purified and identified as diphenylaminobenzamidine hydrochloride. The insoluble residue was crystallised repeatedly from alcohol and obtained in pure white needles melting at 238° (246° corr.). The substance contained no chlorine and formed no salts. (Found: C=73 46; H=6·04· N=13.3. C18H12ON2 requires C=736; H=5·7; N=13-2 per cent.) The solubility and general relations indicated s-diphenylcarbamide.