COHEN AND BRITTAIN: ACTION OF ALKALIS ON AMIDES. 157 1X.-Action of Alkalis on 14micZe.s. By JULIUS B. COHEN, Yh.D., and CHARLES EDWARD BRITTAIN, B.Sc., The Torkshire College. COHEN and ARCHDEACON (Trans., 1896, 69, 91) have shown that many amides of aromatic bases form additive compounds with sodium alco- holates containing 1 mol. of amide in combination with 1 mol. of the alcoholate. That analogous compounds might exist containing sodium or potas- sium hydroxide in place of the alcoholate mas a natural inference, but every attempt to prepare them by t h e usual method employed in the case of the alcoholates mas unsuccessful. I n the latter case, the addition of the alcoholate to the amide suspended in ether, as a rule, yielded a clear solution, which, after a short time, deposited the crystal- line additive compound.If powdered caustic soda is added to acetani- lide dissolved in ether, the alkali remains undissolved, and on filtering and evaporating the solution, the unchanged amide is deposited. An attempt to precipitate the additive compound by adding a concentrated alcoholic solution of the alkali to the ethereal solution of the amide proved equally unsuccessful, the mixture remaining perfectly clear, and on evaporation a t the ordinary temperature depositing a flocculent pre. cipitate, which did not invite further investigation. We then adopted mother method which had been found useful in the preparation of alcoholates where the amide was only slightly soluble in ether. I n this case, the powdered sodium alcoholate was added to the powdered amide suspended in ether, and, after shaking well, the liquid was filtered ; on standing, the additive compound crystallised from the clear solution. During the process of shaking caustic soda with acet- anilide in ether, a very noticeable change was observed.Both powdered caustic soda and acetanilide alone, after being shaken up in cther, rapidly subside; but when the two are mixed together, a bulky, light, and apparently homogeneous powder is produced quite distinct in charac- ter from either constituent. On filtering, a small quantity of a crys-158 COHEN AND BRITTAIN: ACTION OF ALKALIS ON AMIDES. talline deposit appeared in the ether, which, when decomposed with water, gave a strongly alkaline reaction, but the analytical results oh- tained with different preparations did not agree, and the quantity of alkali was invariably too low.Exactly the same thing occurred in the case of paracetotol uidide. I n spite of these indefinite results, the existence of additive coin- pounds seemed sufficiently clearly indicated to justify further experi- ments. It appeared probable that a n amide more soluble in ether than either acetanilide or paracetotoluidide might be more suitable for the purpose, and this view has proved to be correct. We first selected paracetobromotoluidide, as it is comparatively soluble in ether, and, although the yield was small, we obtained very satisfac- tory results by a method similar to that just described. After repeated trials, with the object of improving the yield by the use of different solvents, we returned .to the original method, which, with slight modi- fications, has been adopted throughout.An excess (about 1 gram) of clean caustic soda is placed iu a mortar under a layer of dry ether and finely powdered; 1 gram of the amide in powder is then added and well mixed with the alkali for a few moments. The sodium hydroxide compound, like the alcoholate com- pound, first dissolves and then rapidly crystallises out. By selecting the moment at which solution occurs, and filtering before the new com- pound has time t o separate, a clear solution is obtained which immedi- ately begins t o deposit crystals. The mixture is filtered into a weighed flask, the ether is decanted a s soon as the crystals have separated, and the latter, after being washed once or twice with ether by decantation, are dried in a vacuum and weighed.The substance is then decomposed by water and the amount of alkali determined by titration with deci- normal hydrochloric or oxalic acid solution. I n all cases, well-defined crystals were obtained, and occasionally transparent needles a quarter of an inch long. These compounds exhibit considerable differences in solability in ether. Thus the sodium hydroxide compound of acetanilide, paraceto- toluidide, and a-acetonaphthalide are nearly insoluble, whereas those obtained from ortho- and para-acetobromotoluidide and potassium hydroxide only deposit the additive compound on concentrating the ethereal solution ; apparently the potassium hydroxide compounds are much more soluble than the corresponding sodium hydroxide compounds.They are all decomposed by water, or more or less rapidly in contact with moist air, and, like the corresponding alcoholates, may be dis- sociated, not only i n boiling ether, but even in some cases by washing with cold ether; this occurs notably in the case of the bromine deriva- tives of the amides, probably by reason of their greater solubility, whereas the amides themselves are not affected in this way.COHEN AND BRITTAIN: ACTION OF ALKALIS ON AMIDES. 159 The following amides appear to form additive compounds with caustic soda, but only those have been analysed which gave a satisfactory yield of the pure product. Acetanilide, acetobromanilide, acetiodanilide, paracetotoluidide, par- acetobromotoluidide, orthoacetotoluidide, orthoacetobromotoluidide, a- acetonaphthalide, a-acetobromonaphthalide, P-acetonaphthalide, p-aceto- bromonaphthalide.The potassium hydroxide compounds of ortho- and para-acetobromo- toluidides have also been investigated. It would have been a simple matter to have multiplied examples; but the results obtained convinced us that the reaction was of a general character, and that the compounds were of a perfectly definite type. Apart from a certain theoretical interest which attaches to these compounds and which is discussed further on, they form a class which, we Selieve, has no analogues among inorganic or organic substances, for they may be regarded as containing the sodium and potassium hydroxide in the loose form of combination which is exhibited by water, alcohol, or benzene of crystallisation. Another point of interest is the fact that, as the alkali readily dis- solves in ether in presence of certain of the nmides, and as the former retains its alkaline character unchanged in this solut,ion, an ethereal solution of caustic alkali is thereby obtained, which may be found ap- plicable as a reagent where aqueous or alcoholic potash or soda do not fulfil the requirements of the reaction.We have, for example, at- tempted to prepare glycol from ethylenic bromide in this manner. On boiling up an ethereal solution of paracetobromotoluidide potassium hydroxide with rather more than the calculated quantity of ethylenic bromide for several hours, potassium bromide separated, and the liquid became neutral ; the potassium bromide, after being collected and care- fully washed with ether, was extracted with a small quantity of pro- pylic alcohol in the cold, and the alcohol filtered and evaporated, when a small quantity of a viscid liquid was left.As the ethereal solution might also contain a little glycol, i t was shaken up with water ; on evaporating the water, a few crystals of acetobromotoluidide separated (m. p. 11S0), together with some globules of liquid. The latter were separated by again extracting with water and evaporating, but the liquid thus obtained was so small in quantity that it could not be further examined, although it is not improbable that the substance is glycol. By the action of chloroform on boiling ethereal solution of ortho- acetobromotoluidide sodium hydroxide, the liquid became neutrbl, and sodium chloride mixed with a crystalline compound separated.There was a faint smell of isocyanide, but no formic acid was formed. The compound which separated, along with the common salt, was a t first160 COHEN AND BRITTAIN : ACTION OF ALKALIS ON AMIDES. thought t o be acetobromotoluidide as i t had the same melting point, but it is much less soluble in ether, and insoluble in boiling water; moreover, it crystallises in feathery tufts aod not in needles. We briefly mention these facts, as some little time may elapse before it will be possible t o continue this investigation. Paracetobromanilide sodium hydroxide, C,H4Br *NH* C,H,O,NaOH. -One gram of finely powdered acetobromanilide was added to an excess of caustic soda finely ground under ether, and the whole well mixed. The ether was then filtered ihto a weighed flask, corked, and allowed t o stand 12 hours; after decanting the ether from the crystalline deposit, the latter was dried in a vacuum, weighed, and analysed.0.183 gram required 6.8 C.C. of N/lO oxalic acid. Na = 8.55 per cent. C,H,Br*NH*C,H,O,NaOH requires Na = 9.0 per cent. Pas*cLcetobromotoZuidide sodium hydroxide, CH3*C6H3*Br*NH*C,H,0,NaOH, was prepared as above described, but crystallised from the ethereal solution more readily than the corresponding anilide. The following results were obtained. I. 0.192 gram required 7-5 C.C. N/10 hydrochloric acid. Na = 8.9. 11. 0.214 ,, 7 , 8.1 9 , 9 , Na = 8.7. 111. 0.115 ,, 9 9 4.6 9 , 9 9 Na = 9.0. IV. 0.124 ,, 9 , 4.7 9 , > 9 Na = 8.7.The average on the four determinations is Na=8.8 per cent. CH3* C,H,Br*NH* C,H30,NaOH requires Na = 8.6 per cent. Pnracetobromotoluidide potassium hydroxide, CH,. C6H3Br*NH* C,H,O,KOH, is very soluble in ether, for the whole of the amide in combination with the potassium hydroxide dissolves. One gram of the acetobromo- toluidide was ground up with an excess of potash under ether, filtered, and the ether evaporated in a vacuum ; the residue weighed 1.223 grams (calculated 1-24 grams). It is a comparatively stable com- pound, and requires t o be boiled with water some time before it is completely decomposed ; the residue on titration required 38 C.C. N/10 oxalic acid; K=12*1 per cent. (calculated 13.7). A purer product was prepared by partially evaporating the ether, allowing a portion of the compound t o crystallise, and draining this on a porous plate.The following results were then obtained. I. 0.401 gram required 14 C.C. N/lO oxalic acid. K = 13.6 per cent. 11. 0.3355 ,, 9 9 11.6 9 7 9 , K=13*5 ,, CH,*C,H,Br*NH*C,H,O,KOH requires K = 13.7 per cent,COHEN AND BRITTAIN: ACTION OF ALKALIS ON AMIDES. 161 Orthoacetotoluidide sodium hydroxide, CH3*C6H4*NH*C2H,0,NaOH, which was prepared in the usual way, on titration with N/10 oxalic acid, gave Na = 12.2 per cent, in each of two experiments, CH,*C,H,*NH*C,H,O,NaOH requires N a = 12-17 per cent, CH3*C~H,Br*NH*C2H,0,NaOH, is very soluble in ether, and considerable difficulty was experienced in obtaining a pure preparation, When the ethereal solution was eva- porated, some of the free toluidide crystallised at the same time, and the percentage of sodium was generally about 1 per cent.too low. A.ttempt,s t o precipitate the compound by adding to the ethereal solu- tion indifferent solvents such as chloroform or benzene proved fruit- less. The following melhod was finally adopted ; the sodium hydroxide was powdered under about 10 C.C. of ether, transferred to a flask, and the acetobromotoluidide added in quantities of about 0.25 gram, being well shaken after each addition. The toluidide dissolved readily a t first, more slowly after 0.75 gram had been added, and very soon crystallisation of the sodium hydroxide compound was observed. At this point, the liquid was rapidly filtered, and left to crystallise.Orthoaceto bromotoluidide sodium hydroxide, 0.089 gram required 3.45 C.C. N/10 oxalic acid. N R = 8.9 per cent, CH,*C6H,Br*NH*C2H,0,NaOH requires Na = 8.6 per cent, CH,*C,H,Br *NH* C,H,O,KOH, like the para-compound, is exceedingly soluble in ether, but differs from the latter in forming a non-crystalline glassy mass on evaporating the ethereal solution ; it was therefore impossible to purify the product. Like the para-compound also, it is only slowly decomposed by water, and requires to be boiled with it for some time before complete decom- position is effected. A rough determination of the composition of the substance was effected by shaking up 0.25 gram of the amide with an excess of potash in ether, filtering, evaporating the ether, and titrating the product.It required 9.85 C.C. N/10 oxalic acid. K=12.4 per cent. (calculated 13.7 per cent.). The low resnlt is no doubt due to presence of uncombined amide. a-Acetonaphthalide sodium hydroxide, C,oH7*NH*C,H,0,NaOH, crys- tallises very rapidly, and the ethereal solution must be filtered as quickly as possible after mixing the amide with the alkali. A mean of four titrations with N/10 hydrochloric acid gave Na= 10.23. CloH7*NH*C2H,0,NaOH requires N a = 10.3 per cent. O~thoacetobromototuidide potassium hydroxide, a-Acetobronaonu~ht~~ulicle sodium hydroxide, VOL, LXXIH, M C',,H,Br * NH*C,H,O,NaOH,162 COHEN AND BRITTAIN : ACTION OF ALKALIS ON AMIDES. is fairly soluble in ether and crystallises slowly on standing. A mean of three titrations with N/10 oxalic acid gave Na=’i*5 per cent.CloHGBr*NH* C2H30,NaOH requires Nx = 7.5 per cent. P-8cetonc~pl~tl~alide sodium hythoxide, C,,,H7*NH*C,H,0,NaOH.-The mean of four titrations with NjlO oxalic acid gave N ~ L = 10.2 per cent. C,oH7-NH-C12H30,NaOH requires Na = 10.2 per cent. P-Acetob*om.onaphthalide sodium hydroxide, C,oHGBr*NH*C,H30,NaOH, dissolves easily in ether, but does not crystallise readily, and a pure product could not be obtained. On account of the ready solubility of the potash compound of paracet.obromoto1uidide in ether, it was chosen in order to study the action of various reagents on it. When iodine dissolved in ether is allowed to drop into an ethereal solution of the potash compound i t is immediately decolorised, and a crystalline mixture of potassium iodide and iodate separates.The action of iodine on the substance is there- fore the same as on aqueous potash. A small quantity of iodoform was also detected ; and this, which was also observed in the case of acetanilide sodium methoxide, is probably due in both cases to the decomposition of the ether in presence of the alkali. By the action of acetic chloride in slight excess in the cold, both potassium chloride and potassium acetate mere formed, so that both free acetic acid and free hydrochloric acid must be produced at the same time. Benzoic chloride acts similarly ; the theoretical quantity of benzoic chloride yielded a mixture of potassium chloride and benzoate, together with free hydrochloric and benzoic acids. 1. CH,.C,H,Br*NH*C,H,O,KOH + CGH,*COCl 2. CH,.C,H,Br*NH*C,H,O,KOH + CGH,*COCl Thus : = CH,-C6H3Br*NH*C,H,0 + C,H,*COOH + KCl. = CH,*C6H,Br*NH*C,H,0 + HC1+ C,H,*COOK. The theoretical bearings of these compounds on the constitution of the amides has been discussed in the paper by Cohen and Archdeacon (Zoc. cit.) ; what has been stated there applies with equal force in the present case. If we are to regard these substances as anything more than ‘ molecular ’ compounds, the following formula must be assigned R‘*y*Na for there can be no doubt as to the strict to CH,*C(OH)2’ analogy which exists between them and compounds of the amides with sodium alcoholate. This constitution of the sodium alcoholate compound has recently been disputed by Hantzsch (Annulen, 1897, 296, p. 61) who prefers theFORMATIOX OF MONOMETHYLANILINE FROM DIMETHYLANILINE. 163 R*NH following formula, 1 ONa , but this view can scarcely be CH3' '<OC,H, reconciled with the fact that, by heating the substance on the water bath, it loses a molecule of alcohol and yields sodium acetanilide (Seifert, Bey., 1885, 18, 1358), and that the latter, by the action of methylic iodide, can be readily converted into methylacetanilide and finally into methylaniline. Hantzsch's formula would necessitate a molecular change of a very complex character, which is scarcely justified by the facts.