7 II.-The Alkuline Reduction of :~~~ta.rLit)’u)?iline. By RBPHAEL MELDOLA, l!’..R.s., and ERNEST R. ASDREWS. THE azoxy- and azo-compounds formed by the alkaline reduction of the nitrsnilines and their homologues have for some time figured i n patent specifications as sources of polyazo-colouring inat ters, but hitherto little attention has been bestowed on their purely chemical characters. We have for some time past had under investigation the products of the alkaline reduction of metanitraniline, an examination of which was originally undertaken with a view of finding a, con- venient source of meta-azo-compounds.* Various reducing agents were tried under different conditions, and finally sodium stannite was selected as the most conveuient for carrying the reduction to the first stage.We learnt soon after commencing our experiments that Mr. Arthur G. Green had also been working a t the subject, and he was qood enough to commuuicnte to us the particulars of his method of performing the reduction as well as to place a t our disposal speci- mens of the products which he had obtained. For this information, which has been found of great use, we desire a t once to express our thanks to Mr. Green. Dimetadiam idoaeox ybenzene. I n order to prepare this compound, metsnitzmiline (15 grams) is dissolved in boiling water (1 lit,re) and a cold solution of sodium stannite (40 grams pure stannous chloride crystals dissolved in 100 C.C. cold water and mixed with 40 grams of solid sodium hydroxide dissolved i n 300 C.C. cold water) is slowly added to the boiling solu- tion.On allowing the contents of the flask to cool, the azoxy-com- pound separates as a yellow substance, crystallising in needles. In order to purify the compound, i t is collected, washed with water, dis- solved in dilute hydrochloric acid, filtered to remove tarry matter, and the base reprecipitated by ammonia, this operation being re- pented till the substance dissolves in acid without leaving any residue. Further purification can be effected by crystallisation from toluene, dilute alcohol, or boiling water. From toluene, the substance sepa- rntes in large, golden scales, from dilute alcohol in needles, and from water, in which it dissolves only at the boiling temperature, and then * Concurrently with the above reeearch, I had commenced some experiments 011 the production of meta-azo-compounds by the action of nitrosobenzene on suitable ainido-compomids.The work was interrupted owing to the summer vacation, and Mr. Charles Mills has since entered t,he saine field, and has obtaired results similay to mine (Trans., 1895, 67, 925), so that this part of the investigation is omitted jrom the present communication.-R. M.8 MELDOLA AND ANDREW3 : but sparingly, in very slender, yallow needles. The melting point is 14G-148". A very large number of analyses of various preparations have served t o convince us that the alkaline reduction of metani- traniline is accompanied by the formation of products other than the simple azo- and azoxy-compounds. The azo-compound is, no doubt, simultaneously formed in varying quantities, and its production is unavoidable unless the proportions of sodium stnnnite and metani- trnniline mentioned are rigidly adhered to : but, i n addition, other compounds containing a higher percentage of carbon are also appa- rently produced, and become associated with the basic products.We liavc not been able to isolate any othcr compound a t present, but thc great difficulty which we have experienced i n obtaining specimexis giving correct numbers on analysis has convinced us of the existence of such secondary products. The following resnlts were given by two of the purest preparations. 0.1098 gave 0-2554 CO, a,nd 0.0534 H,O. 0.1216 ,, 24.5 C.C. moist nitrogen at 10" and 765 mm. N = 24.26. 0.1610 ,, 34.3 ,, ,, 15.5' and 755% mm. N = 24.71.C,,H,,N,O requires C = 63.15; H = 5.26; N = 24.56 per cent. The constitution of the compound is expressed by one of the two C = 63.38 ; H = 5.37. formub- I t is a strong base dissolving readily in all mineral acids and in the stronger organic acids ; most characteristic is the insolubility of the dibydrochloride, ClzHlzN40,2HC1, in excess of hydrochloric acid, the salt being at once precipitated in the form of a, whitish, crystalline powder on adding strong acid to the solution of the base in dilute acid. The base dissolres more or less readily in all the usual or- ganic solvents, its solutions being orange ; i t is n o t volatile in steam to a sufficient extent t o enable it to be purified by this means. Of its derivatives, Mixter has prepared the dibenzoyl compound by re- ducing benzoylmetnnitranilide with zinc and alcoholic ammonia ( R m t ~ .Chem. J., 1883, 5, 5 ) , and gives the melting point as about 272". We have prepared the diacetyl derivative by boiling the base for about half an hour with acetic anhydride in acetic acid solution, and crgstallising the product from boiling glacial acetic acid, in which it dissolves but sparingly. I t forms an ochreous powder consisting of micro-crystalline nodules melting at 254". 0.1161 gave 0.2626 CO, and 0.0566 H,O. C = 61.67 ; H = 5.34. 09638 ,, 39.3 C.C. moist nitrogen at 11" and 766.7 mm. N = 17-89.THE ALKALINE REDUCTION OF METANITRANILINE. 9 The bisazimide was also prepared by diazotising i n hydrochloric acid solution, precipitating the diazoperbromide and treating the latter with ammonia (Meldola and Hawkins, Proc., 8, 133).After crjstallisation from petroleum, it was obtaineci in the form of nodular, oclireous crysta,ls, which become darker on exposure to light. 0,1792 gave 62.1 C.C. moist nitrogen at 15.5' and 753.5 mm. N = 40.09. 0.0888 ,, 29.9 ,, ,, ,, 13' ,, 763.1 ,, N = 39.88. N* C ,H4*N3 requires N = 40.0 per cent. O<k.C,H,.N, This compound melts at 85-86', and explodes when heated in a dry tube above its melting point. The yield is not good, as the bisazimide is ccntaminated with a considerable quantity of i l resinous compound, from which it must be separated by extraction with hot, dilute :dcohol, preferably with the addition of animal charwal, before it is in a condition for crystallisation from petroleum.Thc diamidoazoxg-compound is very readily diazotised by the usual methods, and the tetrazo-salts enter at once into combination with amines and phenols to form colouring matters. As a typical azo-compound, that, produced by combination with P-naphthol mas prepared for analysis. The tetrazo-chloride is prepared in the usual way, and the solution mixed with the necessary quantity of p-naph- thol, freshly precipitated by acid from its solution in alkali and suspended as a pulp in the liquid ; uo colour appears at first, but on gradually making the mixture alkaline wilh ammonia, a red tint develops, and, after some hours, st brilliant red precipitate of the azo- colouring matter is formed. The substance, after being collected, washed, and dried, mas purified by repeated cry stallisation from boilicg aniline ; it then formed small, dull, red needles melting at 244-245O.0.1167 gave 15.3 C.C. moist nitrogen at 21.4" and i64.2 mm. N = 15-67. 0.1104 ,, 14.5 ,, ,, 12.75" ,, 766.1 ,, N = 15.64. 0.1015 ,, 0.2665 CO, and 0.0423 H,O. C = 71.GI ; H = 4.63. H = 4.09 per cent. The azo-compound dissolves in strong sulphuric acid with a magenta-red colour, and is precipitated unchanged on dilution with water. I n order to connect the diamidoazoxybenzene with a known com- pound of the same type prepared by another method, a specimen was diazotked, and the amido-groups replaced by iodine by treaking the tetrazo-sulphate with potassium iodide. Much resin was formed,10 MELDOLA AND -4XDREWS : b u t the product, when extracted with alcohol, furnished a crystalline compound consisting of ochreous needles melting a t 118-119*, and having the composition of diiodoazoxybenzeiie. Oa117O gave 6.7 C.C.moist nitrogen at 18' and 753.4 mm. N = 6.61. 0.1426 ,, 0.1476 AgI. I = 55.92. - N-C6H41 requires N = 6%; I = 56.35 per cent. '<kgc6H41 This compound is probably identical with that obtained by Gabi-iel by reducing nietaiodonitrobenzene with alcoholic potassium hydroxide (Ber., 1876, 9, 1408). Di*metadiam idoazo benzene. After many experimeuts with an increased quantity of sodium stannite under various conditions, we finally abandoned this reducing agent as a means of passing from the aeoxy- to the azo-compound. It has been found more advantageous to use for this purpose zinc dust and sodium hydroxide in the following way.The azoxr-compound is dissolved in a small quantity of alcohol, and an equal weight of sodium hydroxide dissolved in a little water is added to the alcoholic solution contained in a flask. A quantity of zinc dust, about twice the weight of the azoxy-compound, is then added, and the contents of the flask kept warm on a water bath for some hours, with frequent agitation ; the solution gradually becomes decolorised, owing to the formation of a hydrazo-compound ; when this stage is. reached, the solution is a t once filtered to remove the excess of zinc, and on exposure to the sir for a few hours the hydrazo- compound is completely oxidised and the uzo-compound regenerated, as indicated by the orange colour of the solution. By this means only can the complete reduction of the azoxy-compound be insured.Potassium hydroxide cannot be conveniently used, as this carries the reduction too far, the yield of azo-compouud being small and metrt- phenylenediamine being formed. The purification of the azo-compound is best effected by crystalli- sation from water ; for crystallisation from organic solvents, especially hydrocarbons, has a tendency to render the compound i m p r e , as it appears to be capable of entering into combination with and obsti- nately retaining some of the solvent. The alcoholic solution contain- ing the aeo-compound, prepared in the manner described, is trans- ferred to a flask, diluted with water, and all the alcohol distilled off; the orange-coloured, crystalline residue is collected, washed with cold water, and dissolved in a large volume of water by means of a little hydrochloric acid.The solution of the hydrochloride is then raised to the boiling point, made alkaline with amnionia, and filteredTHE ALKALINE REDUCTlON OF METANITRASILISE. 11 rapidly while still hot; on cooling, the azo-compound separates in dull, orange-coloured needles. This method of purification is more effective than the direct solution of the azo-compound in boiling water, as the crystals wben once formed dissolve with great difli- cnlty; the treatment can be repeated till t.he hot solut<ion passes tbrough the filter without leaving any residue. The readiness with which the azo-compound combines with other substances renders it a matter of considerable difficulty to prepare specimens sufficiently pure for analysis.Until the above method of preparation and puri- fication was adopted, OUT products all gave a percentage of nitrogen below that required by the formula of the azo-compound, and for some time we were under the impression that the treatment of the azoxy- compound with zinc dust and alkali had not removed oxygen, but had effected transformation into an isonieride, possibly stereochcmical. The melting point (after drying a t 110-120°) is 150-151' when pure. 0.0962 gave 21.45 C.C. moist nitrogen a t 14.2" and 768.7 mm. N = 26.27. 0.0479 ,, 10.9 ,, ,, 20' ,, 768.3 ,, N = 26.27. V*CsH,*NH, K\'.C6Hd*NHZ requires N = 26.41 per cent. In order to characterise this compound further, the bisazimide was prepared by the usual method, but the yield is extremely small, and the melting point the same as that of the bisazimide obtained from the azoxy-compound (86"), so the product was not further examined.The diacetyl derivative was prepared by dissolving the base in glacial acetic acid mixed with acetic anhydride, and heating at 100' for an hour. The product was purified by crystallisation from glacial acetic acid, in which it dissolves somewhat more readily than the cor- responding diacetylazoxy -compound, but as the analytical results still indicated that an impurity was present, it was finally obtained pure by crystallisation from boiling aniline. 0.0952 gave 14-15 C.C. moist nitrogen a t 20Oand 759.5 mm. N = 18-96. 0.1992 ,, 0.4753 CO, and 0.0959 H,O. C = 65.08. H = 5.34. 0.0865 ,, 14.5 ,, 9 9 ,, 21° ), 759.4 ,, N = 19.05.requires N = 18-92 ; C = 64.86 ; H = 5.40 p. c. E*C6H*.NH*C,H,O N* C GH,*NH*CoH30 This acetyl derivative is rery similar in appearance and properties to that obtained from the diarnidoazoxy-compound; it forms a micro-crystalline powder (small needles from aniline) of a somewhat more orange colour than the latter acetyl deriviit8ive. It begins to shrink at 268", and melts completely at 272". The dibenzoyl derivative was prepared hg snspending the crystal-12 RIELDOLA AXD ANDREW'S : line base in warm water, adding the calculated qnantity of benzoyl chloride, and then agitating briskly with the gradual additioa of sodium hydroxide till the solution had become slightly alkaline. The prodnct was collected, washed, and dried, and then crgstallised from boiling aniline.It forms straw-coloured, microscopic needles, melting at 284-285". 0.1587 gave 16.35 C.C. moist nitrogen at 20.5' and 756 mm. N = 13.36. - - R* C 6HPaNH*C7H6O N* C6H,*NH*C7H,O requires N = 13.33 per cent. The diamidoazo-compound is converted by nitrous acid, i n the presence of hydrochloric acid, into w teti-azo-chloride, with the same facility t h a t the diamidoazoxy-compound undergoes this transforma- tion. Combination with [j-naphthol was effected in the same manner as with the previous compound. The trisazo-compound, after crys- tallisation from aniline, consists of small, red needles, duller in shade than the corresponding bisazo-azoxy-compound. 0.1405 gave 19 C.C. moist nitrogen a t 10.3" and 75-53 mm.N = 16.05. 0.0881 ,, 0.2380 CO, and 0.0395 H30. C = 73.68; H = 4-98. H = 4.21 per cent. The substance meIts at 282O, and dissolves in strong sulphuric acid with a magenta colour, which is perceptibly bluer than the colonr given by the corresponding azoxy-compound, the difference bet ween the two becoming more pronounced on diluting with a little water. Like all the azo-derivatives of /3-naphthol, it is destitute of phenolic characters, being quite insoluble i n boiling aqueous alkali. It dis- solves in boiling alcoholic potash with a red coloration, which is indistinguishable from that given by the azosy-compound under the same conditions. Of the salts of dimetadiamidoazobenzene, the oxalate is very cha- mcteristic. I t is formed by adding a solution of oxalic acid to a boiling aqueous solution of the base; the salt separates at once in the form of ochreous scales, which are but slightly soluble, even in hot water.The dry salt has no definite melting point, b u t chars at 205--210'. A specimen dried at ordinary temperatures over strong sulphuric acid i n a desiccator for some days was analysed with the f ol lo wing results . 0.1502 gave 0,3050 COz and 0.0626 H20. 0.1064 ,, C = 55.38 ; H = 4.63. 16.5 C.C. moist nitrogen a t 15Oand 765 mm. N =: 18.29.THE ALRALIXE REDUCTION O F 3IETA"IRASlLINE. 13 From its mode of formation, the diamidoazo-compound may be represented as having the constitution This was confirmed by displacing the NH,-groups by iodine by means of the diazo-reaction. The diiodoazobenzene crystnllises ji; orange scales melting at 150-151° (Gabriel, B e y ., l 8 i 6 , 9, 1410), and is identical with the compound produced by the alkaline reduc- tion of metauitroiodabenzene. 0.0718 gave 4 C.C. moist nitrogen at 10.5O and '758.3 mm. N = 6.61. C,H41*N,*C6H41 requires N = 6.45 per cent. As tlie para-position with respect to the azo-group is open in both benzene rings in the above formula, it appeared of interest t o t1-S whether one molecnle of the compound would combine 1vit.h two molecules of a diazo-salt, so as to form a tertiaryazo-compound of the It was found, liomever, that the base did not combine with diazo- benzene chloride, so a more acid diazo-salt, namely, paranitrodiaao- benzene chloride was used ; combination took place with this diazo- salt in the pi*esence of sodium acetate, and a brown, amorphous substance separated.The latter has all the properties of a diazoamide and not of an azo-compound ; it dissolves readily in alcoholic sodium hydroxide in the cold with rz magenta colour, and with a similar colour in hot aqueous, caustic soda solution ; i t is reprecipituted from its alkaline solutions as a flocculent, ochreous substance on the addition of acids. The compound is uncrystallisable, and could not therefore be obtained pure; it melts with decomposition a t about 198-200"- A determination of nitrogen gave results agreeing fairly Fell with the formula of the bisdiazoamide. ( p ) NO2*C6H4*N2*NH* C6HioN2*CsH~'NH*N**C6HI*T\TOa ( p ) . We have no doubt that this is the compound formed under the conditions described.Il'emarks on the Beduction cf I\li'tyo-coinpotsnds. By R. 35ELDOL-4. The extreme readiness with which meta- and para-nitranilinc* give azox5- and azo-compounds under the influence of alkaline reducing- * Orthonitraiiiline does not give an azoxy-compound when reduced with sodium stannite under the same conditions as metanitrariiline j it is partly reduced to orthophenylenediamine and partly left unchanged.14 MELDOLA: REMARKS ON THE agents, whilst passing with equal readiness into the respqctive diamines under the influence of acid reducing agents, has always appeared to me one of the most striking examples of the difference in behariour of nitro-compounds, according to the mode of reduction, that is, acid or alkaline. In dealing with the theory of the reduction of nitro-compounds the equations X.NO2 + 3Hz = X*NH, + 2820 can obviously be considered only as expressing the 6nal I-esults of a series of intermediate transformations.Of the nature of these in- termediate stages, nothing was known definitely till E. Hoff mann and Victor Meyer showed that nitro-compound8 of the parafin series gave hydroxylamiue derivatives on reduction with stannous chloride ( Rer., 1891, 24, 3528 ; methylh~droxylamine from iiitrornethane). The remarkable discovery by Bamberger (Ber., 1894, 27, 1347 and l.548), and Wohl (ibid., 1432) that nitrobenzene can be reduced to p-pheuylh~droxylamine by zinc dust and water, may be taken as another proof that the mechanism of the process of reductiou is not Amply yepresented by the withdrawal of oxygen from the uitro- qroup.The remarks which I now 7:enture to offer are t o be regarded in the light of a tentative contribution to the theoretical side of this cinestion. Much more experimental evidence will be required before ;L complete theory of t4he process can be formed, but it appears desir- able to call attention at once to the necessity of remodelling the cxisting crude notions respecting a transformation of such general scientific and technical importance. The following attempt to indi- cate, at least, a possible series of steps may be found sufficiently suggestive to prompt further investigation, even if destined to be displaced by other hypotheses as our knowledge grows. Under all conditions, the first action of the reducing agent may be regarded as being the hgdroxylation of the nitro-group.The derivative of the hypot.hetica1 dihydroxylamine being too un- stable to exist, is reduced by further action, thus- O H H X - N < ~ ~ + = X*N< + H,O .. .. .. . . .. (Ir) OH At low temperatures, 2nd with ,weak reducing agents, tho hydroxyl-REDUCTIOK OF NITRO-COMPOUNDS. 15 amine stage can be retained as a resting stage in certain cases. At high temperatures, a i d with strong reducing agents, the second hydroxyl group is replaced. H H X*N < OH + = X*N<H + H,O .......... (111) This view would represent the ordinary conversion of a nitro- compound into an aniine by the usual zethods; it assumes that the hydroxylamine stage is passed through rapidly and imperceptibly, unless special means are taken for arresting the process at that stage.Under other conditions, of which the presence of alkali appears to be the most favourable, and, in a less degree, the access of atmospheric oxygen (Baurberger, Zoc. cit., 1550), the hydrovylamine derivative undergoes " condensation " with the formation of an azoxy- or an azo-compound. To bring about this result two molecules must coalesce, and frQm general considerations concerning such processes of condensation it seems feasible to represent the change, thug- X*N<; ........................... >N*X = H ? 7 + H,O .... (IV) 09 H - iH + HOi X*N--N*X ___-I - ................ The intermediate compound at khis stage would be a derivative of the (at present) hypothetical hydroxy-hydrazine, H,N*NH*OH. The assumption that such a compound is formed harmonises well with the facts, because both the azoxy- and the azo-type are derivable there- from, the former by oxidation and the latter by a repetition of the " condensing " process.These two possibilities are best considered separately. In the first place, the oxidation of such a compound would give rise to an azoxy-compound ; unattacked nitro-compound would act as oxidising agent, or, when all the nitro-compound is re- duced to Stage I (see equation), atmospheric oxjgeii might be effec- tive, as found by Bamberger. HO X*r'oH + ?>N*X = x'r>O + HO>N*X.. .... (V) X*N*H 0 X*N I n the next place, the further (inner) condensation of the hydroxy- hydrazine derivative would give rise to an azo-compound which, from this point of view, is the inner anhydride of its geuerator- (VI) X*T*OH - X*N X*NH X*N - 1 1 + H20 ................The hypot'hesis thus suggests that azoxy-compoands arise, not by direct reduction of nitro-compounds, but by the oxidation of inter- mediate hydroxy-hydrazines-as long as nnaltered nitro-compound is present the azoxy-compound can be continuously formed. No azo. compound could be formed at this stage because the oxidising action16 REMARKS ON THE REDUCTION OF NITRO-COMPOUNDS. on the hSrlroxy-hSdrazine is greater than the tendency to condense to the azo-stage (Equation V). But, when all the nitro-compound is used up, the formation of azoxy-compound has reached its maximum, and the reducing agent can act directly, while condensation to the am-stage can also take place.The question thus arises as to the action of the reducing agent when the azoxy-stage has reached i t s maximum. It seems to me that the dircct removal OF oxygen is again an improbable explanation. The most likely product is the inter- mediate hydroxy-hydrazine. X-N H X-P;T.OH X*N X*NH r>O + & = .... .. .... . I . (VII) As no oxidising agent is present a t this stage, the azo-condez?sation (Equation VI) can now go on continuously. These suggestions are well worthy, as I venture to think, of being submit$ted to the crucial test of experiment ; certainly they are it1 harmony with the well-known fact that the azo-compounds are only obtained by alkaline reduction, as the last stage (unless the reduction be carried on to the hydrazo-stage) of a protracted operation.It appears desirable to reinvestigate in the light of the present hypothe- rjis the pr0duct.s of the reduction of azoxy-compounds, and also to ascertain whether the exclusion of air during the process of alkalitic reduction would accelerate the production of azo-compounds. These and other collateral points snggested in the course of this theoretical discussion will be dealt with experimentally in the course of the pre- sent session. An alternative series of stages from the hydroxylamine derivative (Equation 11) may also altered nitro-compound fol loving way. be suggested. The oxidising action might be regarded as taking place of un- in the XY-OH X*N X-N-OH X-N = l > O + H,O The azoxy-compounds on tihis view also are anhydrides of inter- mediate compounds, namely, derivatives of dihydroxy-hydrazine. The action o f mild oxidising agents (atmospheric oxygen) does convert phenylhydroxylamine into azoxybenzene, whilst strong oxidisers con- vert i t into nitrosobenzene (Bamberger and Wohl). Tn harmony, likewise, with this view is the fact that bydroxylamine itself, under the influence of oxidising agents, evolves nitrous oxide, a decom- position which mamy be brought into line with the above hypothesis.THE CHEMXSTLiY OF DlBROJIOPROPTLT~IOCARBI~ilDE. 1 7 The p~ototype of the azoxy-compounds is thus nitrous oxide or its From this point, ~lihydro-clel.irstive, wliicli is a t present unknown. of view tho generally accepted formula of the azoxy-compounds is correspondi ug to St recker's preferable to the Eormula X*?p*X, 0 formula f o r the dinzoainides. Another point, in favour of some such view of the course of reduction of nitro-compounds, as is here advo- cated, is that it gives promincnce to the analogy with thc process O E d u c t i o n of nitroso- and isonitroso-coinpounds. ,<,;.OH 3 x<oH XH-OH - .x<;:p It is unncccssary t,o elaborate these suggestions froin the specula- tive side in greater detail at present, but it will appear from the foregoing that such an apparently simple change as the conversion of a nitro-compound into an amine is by no means so well understood as the crude equations generally given wonld lead us to suppose.