COHEN ON DTBENZANILIDE. 67 V I I L - N o t e om Dibenxnnilide. By J. €5. COHEB, Ph.D., Owens College, nilanchester. THE researches of Paal and Otten (Ber., 23, 2587), and of Pictet (Ber., 23, 3011), which have recently appeared, induce me to publish the following observations, which 1 made some time ago. I n the latter portion, I was greatly assisted by Mr. F. Brownsword, B.Sc., a former stxdent in this laboratory. When a mixture of phenjlthiocarhimido with benzaldehyde is heated for some houys at 190", sulphuretted hydrogen is evolved, and several products are formed. These coiisist mainly of benzanilide and benzaldehyde-aniline, and form a semi-solid mass from which the benzanilide may be isolated by filtration. The latter was recrystal- lised from alcohol and melted at 161-162".When heated in a sealed tube with concentrated hydrochloric acid, it splits up into aniline and benzoic acid. The following result was obtained on analysis :- 0.1995 gram gave 12.4 C.C. of N at 16.5" and 761.2 mm. Pound. Calculated. N.. .... .. 7.36 7-10 The reaction is a, complex one, and was not studied further. A similar reaction described by Losanitsch (Bey., 6, 17C;), and repeated by Higgin (Trans., 1882, 132), consists in heating phenylthiocarbimide with benzoic acid at 220". A solid product melting at 160" is thus formed, which is stated t o be dibenzunilide; but in neither paper is any analysis given. The similarity of the two reactions, and the alleged difference in the product obtained, induced me to repeat Losanitsch's experiment. The product, of the reaction is a solid, crystalline mass, which, wheu recrystallised f r o m alcohol, melted at 161-162".The following result was obtained on analysis :- 0.2150 gram gave 13.4 C.C. of N at 13.4" a d 7'75.5 mm.68 COHEK ON DTBENZXNILIDE. Calculated for Calculated f o r Foiind. benzanilide. dibenzanilide. N.. ...... 7.60 per cent. 7-10 4.63 The compound is obviousiy beuzanilide. 'In t h i s as in all siibsequent amalgses, the nitrogen nlnne has heen determined, aifi this suffices to establish the identity of the compound with benzanilide or dibeiizanilide. To confirm the above resnlt, and t o compare the product with dibenzanilide, I attempted to prepare the latter by the method of Gerhardt and Chiozxa (Ann. Chim. P72?ys. [3], 46, 129) by heatinq a mixture of benzanilide and benzoyl chloride in niolecular propor- tion a t 160-180".In doing so, I followed in detail the method of Oerhnrdt and Chiozzn : but was entirely unsuccessful in obtaining dibeuznnilide." The substance was purified by the method given in their paper, namely, by digesting with soqium carbonate solution, washing w i t h water, and recrystallising the product from alcohol. Prepared in this was, the cornpound obtained is a white, granular mass consisting of acrgregates of microscopic crystals, differing entirely in appewance from the glistening plates of benzanilide. The substance, which was evidently impure after two recrgstallisations, melted a t 158", and gave the following results on analysis :- 1. 0.2020 gram gave 13.1 C.C. of N at 24" and '759.9 mm.2. 0.2168 ,, ,? 14.2 ,, ,, 15.5 ,, 747.8 ,, Nitrogen found. 1 ............ 7.54 per cent. 2 ............ 7-64 ,, The above process was then repeated under a variety of different coiidit,ions. The mixtme of benzanilide and benzoyl chloride was heated a t 200", 220°, and 230" f o r different periods from 3 to 12 hours, and also in sealed tubes a t 180". In the latter case, it may be mentioned that no excess of pressure was observed on opeiiing the tubes. This would prove that 110 great * The authors state that gaseous hydrogen chloride is evolred during the re- action, but this is not the case. The evolution of hydrogen chloride is not more apparent than when benzoyl chloride alone is hmtcd a t the same temperalure in an open vessel, as I observed by placing benzoyl chloride in a flask in the same bath with the mixture.There was no marked evolution of gas in either case. Gerhardt and Chiozza obtained, as a result of one analysis, 5 per cent,. of nitrogen in place of 4.6 per cent. The percentage, which I recalculated from their figureq, should be 5-22 per cent. This is undoubtedly due to the fact that the benzanilide, which remains unaltered in the process, contains benzoic acid, from which it, is ex- ceedingly dificrilt to free it. The low melting point (137') may also be accounted for in this way.COHEN ON DIRENZANIL'TDE. 69 evolution of p s had occurred during the heating. I n the many an;llyses made of both the impure and purified product, in not one instance did the percentage of nitrogen fa11 below 7 per cent.The following are some of the results of the analyses of the pro- ducts obtained at 180". 1. 0.2170 gram gave 13.4 C.C. of N a l 36" and 750 mm. 2. 0.2068 ,, ,, 12.65 ,, ), 12.9" ,, 739.85 mm. 3. 0.2160 ,, ,, 13.4 ,, ,, 13.5" ,, 750.9 ,, Per cent. of N found 1. After heating in a sealed tube for 4 hours at 180" and Melting point 2. After heating at 180" for 4- -5 hours and crystallising 3. Aft,er heatiyg at 180" for 4-5 hours and crystallising crystallising the product three times. 158 " .......................................... 7-19 o n c e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.13 t h e e times.. .................................. 7.34 The following experiment was then carried out :-45 gi'ams of pure benzanilide melting at 160-160*5" were heRted at 180" with 34 grams of benzoyl chloride for five hours, and the liquid product poured whilst hot into a large volume of cold water.To prevent possible decomposition of dibenzanilide, t,he product was digest ed with sodiiim carbonate solution in the cold, by adding, successively, small quantities of a solution of sodium carbonate until an alkaline reaction remained perma'nent. It was then filtered, dried, and weighed. 43 grams were obtained. The substance was then sub- mitted to a process of recrystallisation from alcohol, and the melting point of each product determined as follows :- F i ~ s t CrystaZZ1:sation.-Granular mass, m. p. 1.55", also a minute quantity of needle-shaped crystais melting at 156". Second 01.1Jsttallisation.-~hite grains, m.p. 157-158". A few needle-shaped crystals again formed and were carefully separated. Tltese melted at 157". Third CrystaZliPation.-Appearance similar to the previous product, m. p. 157-158". Fourth CrystaZZisatio7z.-Appearance unchanged, m. p. 1.57-159'. No needles. Fifth Cry.FtalZ~.scction.-The crystals were more distinctly tabular., and had the characteristic glistening appearance of benzsnilide, m. p. 157-159'. By evaporating tbe mother liquors from the separate crystallisa- tions, more granular crystals mixed with a brown resinous matter No needles were present.70 COHEP; ON DIBENZANILIDE. separated, and from the last mother liquor, a small amount of benzoic acid was obtained. The presence of benzoic acid is not easy to explain.The amount of brown, resinous matter increases with the temperature at which the reaction is carried out. I n an experiment made at 230", the product was very impure from this cause, and after five crystallisations from alcohol, only melted slowly at 140-151". Under these conditions, the percentage of nitrogen in the compound was always above tlhe calculated amount. The quantity of the needle-shaped crystals was too small to allow me to do more than make a determination of the melting point. I attempted to obtain more of them by evaporating the mother liquors and dropping some of the needles into the solution t o bring about crystallisation. On the assumption that benzanilide might be dimorphous, I also added a crystal or two to a saturated solution of pure benzanilide.I n both cases I was unsuccessful in obtaining the needles. The literature of the subject would be incomplete without reference to the work of Steiner on tribenzhydroxylamine (AnnaZen, 178, 235). The author states that by heating a-tribenzhydroxylamine in a closed tube, dibenzanilide is formed and carbon dioxide evolved. The evidence for this is based on the decomposition of about 1 gram of substance, which gave a volume of gas roughly corresponding to the amount required by the decomposition of the substance in the manner indicated. He admits that the product does not resemble dibenz- anilide, that it does not smell of phenyl cyanate; but possesses a strong odour of bitter almonds, and is to a great, extent soluble in cold ether. The residue insoluble in ether, after cry stallisation, formed small needles with the melting point 161".Steiner, finding that the melting point did not agree with that of Gerhardt and Chiozza's compound, repeated their experiment of heating together benzoyl chloride and benzanilide. The product, after recrystallisation, had the required melting point of 161". The analysis gave 5.38 per cent. of nitrogen in place of 4.63 per cent. Like G-erhardt and Chiozza, Steiner probably obtained benzanilide containing benzoic acid." The author concludes by stating that the products of decomposition of a-tribenzhpdroxylamine " are not wholly, apparently not even to any extent,, dibenzanilide and carbon dioxide." We map then conclude that, up to the present, dibenzanilide has This compound may therefore be dibenzanilide.* This experiment was again repeated in 1882 by Higgin (Trans., 1882, 132), at :I, temperature of 230'. The author obtained needles melting at 13G0, but as he only determined the percentage of carbon in the compound, viz., 79-60 (benzanilide has 79-18 and dibenzanilide 79.73 per cent.), no satisfact,ory conclusions can be drawn.‘iz EASTERFIELD : PHEXYLBROJXACETIC ACID. not been prepared, at any rate not in the pure state. The synthesis of this compound I desire to reserve for future study. Paal and Otten, as well as Pictet, point out in the papers referred, t o a t the beginning of this note, that by the action of bcnzoyl chloride on acetanilide, benzanilide and acetyl chloride are formed in theoretical quantities, and, from this and nther experiments, state that, in this reaction, the acid chloride of higher molecular weight replaces that of lower molecular weight in combination with tho amine. I had occasion incidentally to try and replace the amido-hydrogen atom in benzanilide by acetyl, and attempted this by heating benzanilide with acetic anhydride and a small quantity of fused sodium acetate. The benzanilide was converted into acetanilide which melted a t 112’. It appears, therefore, that under these con- ditions the action is reversed. I intend to t r y this reaction with the ot.her homologues of benzanilide. This is an additional confirmation of the fact that the replacement of the second amido-hydrogen atom by an acid radical in aniline is not readily accomplished by the ordinary methods.