424 RUREMANN AND CARNEGIE THE ACTION OF ACETONE XXXV.-The Action of Acetone 0% Ammonium Salts of Fatty Acids in presence of Dehydraticng Agents. By DF. S. RUHEMANN and D. J. CARNEGIE. SOME years ago Zanoni (Bey. 15 528) investigated the action of phosphoric anhydride on a mixture of acetamide and glycerol and obtained P-picoline as a product of the reaction. Afterwards Hesekiel (Ber. 18 511 and 3091) more thoroughly studied the reaction and found that it involves the previous formation of acraldehyde whilst the acetamide acts merely as a convenient source of ammonia which, combining with the acraldehyde under the influence of the dehydrating agent yields P-picoline in accordance with the following equation :-2C.3H.40 + NH = CsH,N + 2HZO. Hesekiel extended this reaction t o the formztion of the higher members of the pyridine series using aldehydes instead of glycerol.Acetic aldehyde for instance he found to be first transformed by phosphoric anhydride into crotoiiic aldehyde which in turn con-densing with the ammonia derived from the acetamide forms methyl-ethylpyridine. This transformation of the aldehydes of the fatty series into the homologues of pyridine induced us to investigate analogous methods involving ketmes instead of aldehydes. Briefly stated our experimental method consisted in heating mix-tures of acetone and ammonium salts of fatty acids with dehydrating Agents. In o u r first experiments dehydration was effected by heating the mixture with zinc chloride in sealed tubes at a temperature of 250" for about four hours; but the pressure of the gases evolved under these conditions was so great as to cause much inconvenience.I n our later experiments phosphoric anhydride was substituted for zinc chloride and the reaction was allowed to proceed in ordinary flasks fitted with reflux condensers. The charge for each flask was ON AJIYONIUM SALTS OF FATTY ACIDS. 485 30 grams of acetone, 40 grams of ammonium salt, 35 grams of phosphoric anhydride. After 24 hours' heating on the sand-bath the contents of each flask were submitted t o distillation in a current of steam in order to separate the hydrocarbons which are always formed in this reaction ; excess of solid potash was then added t,o each flask care being taken t o keep the temperature low the while and the resulting strongly alkaline liquid subjected to a second distillation with steam.The preater part of the base formed appeared as an oily layer floating on the surface of this second distillate. To liberate that portion of the base which i8 dissolved in the water of the distillate solid potash was added thereto since the base though fairly soluble in water is insoluble in concentrated potash solution. The base was finally dried over solid potash and fractionally distilled through the temperature interval 80-200". Tl e second fraction (161-163') on treatment with hydrochloric acid and platinic chloride gave a platinochloride but at the same time an oil, undoubtedly a hydrocarbon rose to the surface of the liquid. The platinochloride was purified in the usual way by repeated decompo-sition by hydrogen sulphide followed by repeated precipitation with platinic chloride.It was dried at loo" and burnt in a stream of oxygen. As will be seen from the appended numbers the analysis undoubtedly proved it to have the composition (CgH,5N)2,H,PtC1,. The first fraction (80-160") did not give a platinochloride. Theory. Found. P t 28.4 28.3 C . . 31.6 3 1.4 H . 4.68 4.7 N . 4.06 4.07 Each of the fractions (165-167") (1 TO-l80") (180-185"), gave platinum salts which were purified by recrystallisation from water. In all cases the analyses qf these salts gave a mean of 28.3 per cent. of platinum; hence it is clear that the chief product of the reaction of acetone on ammonium acetate in the presence of phos-phoric anhydride is the base CgHI5N ; but that other bases of higher boiling point are also present in small quantity is evident from the behaviour of the liquid during distillation.The mixture of hydrocarboris formed along with the base C,H,,N in this reaction was also examined and was found to consist chiefly of mesitylene (b. p. 165-167') mixed with small quantities of hydro-VOJ,. LIII. 2 426 RUHEMANN AND CARNEGIE THE ACTION OF ACETONE carbons of higher boiling points. The analysis of the fraction 165-167" gave the following result :-Theory for mesitylene. Found. c . 90.0 89.:3 H . 10.0 10.2 I t should in this connection be noted that when zinc chloride is used to dehydrate much more hydrocarbon and much less base is formed than when phosphoric anhydride is used. Experiments similar to the foregoing were now made with ammo-nium formate and ammonium butyrate ; and from the analyses of the platinochlorides we found that the bases formed in each case aye identicaI with the base C,H,,N already prepared from ammonium acetate :-Analyses of Pt salt Analpsis of Pt salt of base from of base from Theory for H*COONH,.C,H,.COONK,. (CgH,,Pi),,H2PtCl,. Pt 28-42 28.6 28.4 The mixture of hydrocarbons formed in the case of ammonium formate undoubtedly consisted largely of mesitylene as before. I n the case of ammonium butyratre the bye-product distilled over almost complet.ely at 110-113". As it could not possibly be mesitylene an analysis was made. The appended numbers show that the substance was not as we had anticipated a hydrocarbon but the ethyl salt of butyric acid -Theory for C,H7.COOC2H,.Found. C 62.0 61.68 H . . 10.3 10.28 The identity of the compound was further proved by the fact that the product of its hydrolysis gave the iodoform reaction. From these experiments i t follows that in all. cases only the ammonia of the salts plajs a part in the formation of the base C9H15N. This base is identical in all its properties with that isolated by Heintz (Annalen, 174 167 and 183 276) from the products of the reaction of ammonia and acetone and which was called by him dehydrotri-acetonamine, Canzoneri and Spica (Ber. 18 51 and 331) obtained it as chief product on heating a mixt'ure qf acetone and acetamide with zinc chloride a t 135-140". They also studied the action of reducing agents on the base and found that it was thus transformed into methylated piperidines.From these observations we may conclude that dehydrotriacetonamine is a hydrogeiiised pyridine-derivative. Accordingly we now tried the action of mild oxiclising agents on th ON XJfMONIUM SALTS O F FATTY ACIDS. 427 base C,H,,N in the hopes that we might by this means dehydrogenise it and so obtain some well-defined member of the pyridine series. Some of the base (boiling a t 161-163') was treated with a filtered solution of bleaching powder ; the mixture became wai-m and a pre-cipitation of calcium hydrate took place. After allowing it to remain for an hour the mixture was heated for some time on the water-bath, and t>hen distilled with steam. In this way we obtained an oil differing from the original base in its very pungent odour and also in the fact that it was heavier than water.We had merely succeeded, however in chlorinating the base not in reducing it. The chlorinc-derivative obtained was exceedinglj unstable ; on attempting to distil it sudden decomposition took place at 80" and chloroform distilled over in large quantity. An analysis of the unpurified substance was made but no conclusions could he drawn from the result viz. 27.7 per cent. C1. Again in the hopes of obtaining a pyridine-derivat'ive we treated the chlorine-derivative with alcoholic potash and then distilled in a current of steam. On saturating the distillate with potash a fair quantity of a light oil was obtained. This was separated dried over solid potash and distilled.It distilled over almost completely between 78" and 79". Three analyses and three vapour-density determinations were made which showed that tlie action of alcoholic potash on the chlorine-derivative does not simply remove hnloid acid, and give a pyridine-derivative as we had hoped but that a more deep-seated decomposition takes place resulting in the formation of large quantities of an isopropyl alcohol hydrate of the formula C3H30,H20. Appended are the results of analysis :-Found. 7 Theory for r-A-C,Hi,O:. I. 11. 111. C 46.15 46-48 46-36 45.91 H 12.82 12.6 12-67 13.7 Theoretical vapour-density of C,H1,O2 referred to hydrogen. Found. 39 39.5 Hydrates of isopropyl alcohol have already been observe& b u t so far as we are aware the known hydrates do not include the one above referred to. Erlenmeyer (Anxaleiz 126 308) describes a hydrate of the formula 2C,H80,H,0 ; whilst Linnenian (AnnaZen 136 40) has isolated hydrates having the compositions 3C3H80,2H20 arid 3C3H80,H20. U?& emity Laboratory Cmn b ridge. 2 G