HARTLEY AND STUART: THE MISCIBILITY, RTC. 309XXXIV.- The Miscibility of Axobenzene and Axoxy-benzene in the Solid State and the SupposedExistence of a Stereoisomeride of Axobenzene.By HAROLD HARTLEY and JOHN MCARTHUR STUART.IN the preparation of azobenzenc by the electrolytic reduction ofnitrobenzene in aqueous alcohol containing sodium acetate witha nickel cathode a small quantity of a deep red liquid was occa-sionally obtained, which crystallised on keeping, giving radiatingmasses of orange, acicular crystals melting a t 25O. It appeared tobe identical with the substance obtained by C. V. and R. A.Gortner (J. Arne?*. Chem. SOC., 1910, 32, 1294) as a by-produc310 HARTLEY AND STUART: THE MISCIBILITY OFin the preparation of azobenzene by the distillation of azoxy-benzene with iron filings.The authors describe the by-product ofthis reaction as forming orange-red, stellate groups of needles,readily soluble in ether, light petroleum, alcohol, acetone, or methylalcohol, and melting a t 25O to a deep red liquid. Analysis gaveN = 15.35, azobenzene requiring N = 15.38 per cent. They say,further: ‘ I It will not crystallise when seeded with the 6 8 O substance(ordinary azobenzene), but in two instances we have succeeded inquantitatively transforrning the 25O substance into the 6 8 O sub-stance, once by boiling with dilute hydrochloric acid, and in theother instance the cause of conversion is unknown to us. We cangive no method which will always ensure conversion, nor havewe been able to trarsform the 6 8 O substance into the 2 5 O modifica-tion.According to Holleman, the syn-modifications are the leaststable, so it seems probable that the new compound is syn-azo-benzene. ”Attempts to discover the conditions which favoured the produc-tion of the substance of low melting point in place of ordinaryazobenzene by the electrolytic method failed, as little of it wasobtained. It was then found that carefully dried azoxybenzenewhen distilled with iron filings gave, as a rule, nothing butordinary azobenzene ; however, on moistening the azoxybenzsnewith alcohol, or, better still, with aniline, a considerable quantityof the product of low melting point was obtained. If aniline wasused, the solution of the distillate in light petroleum was wellwashed with dilute hydrochloric acid before crystallisation.Various specimens gave melting points varying from 25’2O to 26O,and two nitrogen estimations gave 14.95 and 14.84 per cent.respectively, showing that the substance is not pure azobenzene,which contains 15-38 per cent.of nitrogen. The resemblance ofthe crystalline form to that of azoxybenzene, and the fact thatthe drops of the supercooled liquid crystallised if seeded withcrystals of azoxybenzens, suggested that the orange crystals meltingat 25O might be a solid solution of azobenzene in azoxybenzene.The results of determinations of the freezing-point and melting-point curvm of mixtures of those two substances confirmed thisconjecture. The freezing-point curve was determined in a Beck-mann apparatus in the usual way, and the melting-point curve byallowing mixtures of known composition to crystallise slowly, andthen finding the temperature at which melting begins when a thin-walled tube containing the mixture is slowly heated.The melting-point curve determined in this way is necessarily rough, and givesonly an approximate estimate of the composition of the solid phase.The horizontal portion of the curve must pass through the eutectiAZOBENZENE AND AZOXYBENZENE IN THE SOLID STATE, ETC. 311point, but it will be noticed that the experimental points lie fromlo to 3 O above i t owing to the difficulty of deciding when meltinghas really begun to take place.The experimental results are given below :Molecular per-centage of azo-benzene.08.312.918.222.625.929.134.838.142.448.955.162.680.594.1100.0Freezing point.35.031.329.327.125.325.527.833.135.839.244.147.752.260.465-867.9Molecular per-centage of azo-benzene.02.93.66.913.525.234.847.356.964.075.091.5Melting point.35.330.929.627.825.626.726-427.726-834.648.062.1The form of the diagram shows that the substances give two setsof mixed crystals with an eutectic temperature of 24'5O; a t thistemperature crystals of azoxybenzene can dissolve about 10 percent.of azobenzene molecules, whilst those of the azobenzenedissolve about 45 per cent. of molecules of azoxybenzene.The crystalline form of azoxybenzene has not been fully investi-gated.Bodewig (Zeitsch. Kryst. &€in., 1879, 3, 411) measured theprism angle of the needles (87O32/), and from their optical proper-ties described them as orthorhombic. The crystalline form ofazobenzene has been examined by Calderon (Zeitsch. Kryst. Min.,1880, 4, 234), and by Boeris (ibid., 1901, 34, 301), who describesit as monoclinic with a prism angle of 55O3', and axial ratios:a : b : c = 2-1076 : 1 : 1.3312 ; /3 = 114O26'. The evidence is thus insuffi-cient to decide whether there is any relationship between the normalcrystalline structures of the two substances. Their miscibility inthe crystalline state is probably similar to a number of instancesnoticed by Jaeger (Zeitsch. Kryst. Min., 1907, 42, 236), where themiscibility seems to depend, not on a relationship of crystallineform, but cn the similar spacial structure of two molecules, forexample, the miscibility of azobenzene and benzylaniline (MissIsaac, Yroc.Roy. SOC., 1911, A , 84, 344). Crystals produced bythe crystallisation of mixtures of azobenzene and azoxybenzenefrom light petroleum appeared identical in every respect with theproduct of low melting point obtained by the electrolytic reductionof nitrobenzene or by distilling azoxybenzene with iron filings, an312 HARTLEY AND STUART : THE MISCIBILITY, ETC.i t seems certain that the substance obtained by C. V. and R. A.Gortner, and described by them as a new stereoisomeride ofazobenzene, was a mixed crystal of azobenzene and azoxybenzene.Their analysis gave N = 15.35, the theoretical content forazobenzene being N = 15.38 per cent.; the specimens analysedabove gave N=14*95 and 14.84; the eutectic mixture wouldcontain about 14.4 per cent. of nitrogen, so that the substanceanalysed probably contained some azobenzene.The production of the mixed crystals, owing to the contamina-tion of azoberzene by azoxybenzene, is easily explained in bothprocesses where they have been obtained. In the distillation of70603020Molecular perceningc of azobenxcnc.- G - Freezing-point curve. - _ + - - Jlelki?tg-point CZLTV~.azoxybenzene with iron filings the product of low melting pointwas only obtained in quantity when some volatile substance waspresent which carried over the azoxybenzene a t a. temperaturebelow that a t which it decomposes. I n the electrolytic reduction,since azoxybenzene is a product of the reduction of nitrobenzenein alkaline solution, some will be present if the current is passedfor an insufficient time to reduce it further. I n the unsuccessfulattempts t o prepare the substance of low melting point byelectrolytic reduction, on ;he assulnption that it wae a stereo-isomeride of azobenzene, the current was always continued untilreduction was complete.PHYSICAL CHEEMISTKY LASOKATUILY,BALLIOI, COLLEGE AND TlLlNITY COLLEGE, OXFOKD