XLII.-The Isomerism of the Alkyl-derivatives of Mixed Diazoamidocompounds.

By RAPHAEL MELDOLA, F.R.S., and F. W. STREATFEILD, F.I.C.

THE isomerism which has been brought to light in the course of our investigations of the alkyl-diazoamides is, as we have already pointed out in former papers, quite inexplicable by the formula generally assigned to these compounds, and in our last communication on this subject we promised to make this isomerism the subject of further research (Trans., 1888, 664). For the complete appreciation of the nature of the problem which we have taken up, it will be advisable to state once more the general conclusion to which our former work pointed and the principle of the method which we have employed in arriving at this conclusion.

The generalised formula of the diazoamides of mixed type is generally written

X-N, NHY or Y·N, NHX.

According to this view, two alkyl-derivatives only are possible, viz. :

X-N2 NR'Y and Y-N2 NR'X.

When mixed diazoamides are prepared, only one compound can be obtained from any pair of dissimilar amines, Griess having shown that it is immaterial in which order the amines are diazotised and combined, a discovery which has been since confirmed by many investigators. We may add that we have made numerous experiments in the course of our work with many pairs of dissimilar amines, and have invariably obtained the same compound in whichever order the amines were diazotised and combined, so that there can be no doubt as to the generality of Griess's conclusions.

The diazoamides, both normal and mixed, can be readily alkylated by heating them in alcoholic solution with the theoretical proportions of alkyl iodide and potassium hydroxide.

Alkyl-derivatives can also be prepared by acting with a diazotised amine upon an alkylamine, such, for example, as by the action of X-N2 Cl on Y NHR' or of Y N2 Cl on X⚫NHR'. Our first experiments were undertaken with the object of determining the constitution of the mixed diazoamides by comparing their directly alkylated derivatives with the two alkyl-derivatives prepared from the corresponding amine and alkylamine as above. It was found that the derivative prepared by direct alkylation corresponded with neither of

the other isomerides, and it was accordingly inferred that any pair of aromatic amines could give rise to the three isomeric alkyl-diazoamides :

2. By the action of Y⚫N, Cl on X⚫NHR'.

3. By the action of KOH and R'I on X·N,H⚫Y.

The generality of this conclusion has hitherto rested only on the two cases made known in former papers, viz., the methyl- and ethylderivatives of the diazoamides obtained from para- and meta-nitraniline. In the present paper, we give the results of the extension of our experiments to many other pairs of amines, and the truth of our former inference is, as will be seen, completely borne out in every case. From each pair of amines experimented with, we have obtained three isomeric alkyl-derivatives. These series of three isomerides are referred to in the paper as triplets. The compounds produced by Methods 1 and 2 (see above) are easily prepared, and in most instances have presented but little difficulty in the way of purification. The compounds of this class are for the present formulated in accordance with their mode of production on the types X-N,NR'Y and Y⚫N, NR X. The compounds prepared by direct alkylation, which are provisionally formulated on the type X-N,R'Y (Method 3), have been much more difficult to obtain in a state of purity, a difficulty which we attribute entirely to the tendency of the mixed diazoamides to become contaminated with traces of the normal diazoamides by secondary reactions occurring during their formation. We are not aware whether other experimenters have experienced a similar difficulty in dealing with these compounds, and on referring to the various papers by previous workers we find that the ccmposition has, in nearly all cases, been considered as settled by the determination of one element only, generally the nitrogen. It will be seen, however, that this mode of procedure cannot be regarded as satisfactory, since a mixed diazoamide containing a halogen-atom in one radicle might be contaminated with an appreciable quantity of the normal compound without affecting the percentage of nitrogen beyond the ordinary limits of experimental error. To give a specific instance: By the action of paradiazotoluene chloride on parachloraniline, or of parachlorodiazobenzene chloride on paratoluidine, there is formed a mixed diazoamide of the formula C.H,Cl·N2H·C2H1, which contains 17.1 per cent. of nitrogen and 14.8 per cent. of chlorine. The corresponding pair of normal compounds would be

CH,CIN,H-C,H,CI

(containing 26.7 per cent. of Cl and 158 per cent. of N), and

C2H2 N2H·C2H, (containing 18.7 per cent. of N). The two normal compounds differ, therefore, from the mixed compound only by 13 and 16 per cent. of nitrogen respectively, so that a considerable quantity of one or both the normal compounds might be present in admixture with the mixed diazoamide without being detected by an analysis in which the nitrogen was the only element determined.

From considerations such as the foregoing, it is obvious that the purity of the mixed diazoamides can only be regarded as satisfactorily demonstrated when at least two of the constituent elements are determined. The mixed compounds containing a halogenised radicle on one side of the NH-group are particularly well adapted for showing when a pure product is obtained, because the presence of a small quantity of one or the other of the normal compounds is easily shown on determining the halogen; whereas the nitrogen determination would fail to reveal the presence of a small percentage of such an impurity. For this reason we have made use of parachloraniline or parabromaniline as one of the amines in the preparation of all the mixed diazoamides described in the present paper. It is evident that such compounds are more suitable for the purposes of the investigation upon which we are engaged than the more symmetrical mixed compounds containing such radicles as (p)NO, CH, and (m)NO2*CH each side of the NH-group. Thus the mixed compound, (p)NO2 CH ̧•N2H·C ̧H ̧·ÑO2(m), described by us in a former paper, would have the same percentage composition as the two normal compounds,

on

(p)NO2 C&H, N2H·C2H ̧·NO2(p) and (m)NO2 CH1•N2H·C2H ̧•NO2(m), either or both of which might be present in admixture with the mixed parametadinitro-compound without being capable of detection by analysis. It was only after considerable experience with the mixed compound containing one halogenised radicle had shown us that varying quantities of one or the other of the corresponding normal compounds were generally formed at the same time as the mixed diazoamide, that we were led to discredit the purity of those mixed compounds prepared by previous investigators, and taken as pure on the strength of a nitrogen determination only. For the same reason we were led to suspect that our own preparations of the mixed para-metadinitrodiazoamidobenzene might have been contaminated with more or less of the normal dipara- or dimeta-nitrocompound, analysis being obviously of no avail for the decision of this question. We have, therefore, submitted the mixed dinitrodiazoamide from paranitraniline and metanitraniline to a further critical examination, in order to clear up this point in the first place, slight variations in the melting points of different preparations having con

vinced us, in the light of more recent experience, that traces of the normal compound might have been present in our earlier experiments. We shall describe the results of these experiments before passing on to the other triplets. With respect to the melting points of the diazoamido-compounds and their alkyl-derivatives, it will be necessary at this stage of our work to offer some general obser

vations.

Starting from the fact, which we now regard as established on a fairly wide basis, that each pair of amines gives rise to three isomeric alkyl-diazoamides, we have found that the melting point of the compound produced by direct alkylation is always more vague than that of either of its isomerides. This vagueness of melting point appears to be a natural property of the compounds of this series. The explanation of this property is to be found in the fact that the diazoamides, as well as their alkyl-derivatives, are all more or less prone to decompose, with evolution of nitrogen, on heating; so that the melting point is, in most cases, also the point of decomposition, and it is well known that under such circumstances sharpness of melting point cannot be expected. The products formed by direct alkylation, taken as a whole, are more easily decomposable than their isomerides, both on heating and by the action of acids; and it is this fact which accounts for the wider range of their melting points. The instability is more pronounced in those compounds which contain acid radicles, such as NO2, Br, Cl, &c., in both nuclei, and in some of these cases the substance begins to shrink and soften many degrees before fusing to a clear liquid, bubbles of nitrogen being given off more and more freely as the temperature rises. It must be understood that this property more especially pertains to the compounds produced by direct alkylation the isomerides formed by the action of diazotised amines on alkylamines have fairly sharp melting points owing to their being more stable under the influence of heat. We had noticed this vagueness of melting point in the directly alkylated products of the mixed para-metadinitro-diazo-compound first studied, but we were not aware till we had prepared many analogous compounds that the property in question was characteristic of all the alkyl-derivatives produced in a similar manner.

Further Experiments to test the Chemical Individuality of Para-metadinitro-diazoamidobenzene and its Alkyl-derivatives.

Experiment I.-Metanitraniline was diazotised and combined with paranitraniline in the usual manner. The crude product when dry had a melting point of about 200-201°. On crystallising from alcohol, the following melting points were observed:

The melting point after this remained constant, and we believe this to be about the true point of fusion of the substance, the higher readings of the first and second crystallisations being possibly due to the presence of some of the normal compound,

(p)NO2 C2H, N2H·C2H ̧•NO2(p),

which was removed by the continued crystallisations. As illustrating the difficulty of determining the melting point-even in the case of this compound, which fuses much more sharply than its alkyl-derivatives-we may add that the rate at which the sulphuric acid bath is heated materially influences the result. By causing the mercury of the thermometer to rise very rapidly, the melting point of the compound could be raised to 222° or 227—228°.

Experiment II.-In the next experiment, the order of combination was reversed; paranitraniline was diazotised and then combined with metanitraniline as before. The crude product when dry fused at about 210-211°.

After 1st crystallisation: m. p. 210-210.5°

From this result it appears that a purer product was obtained in the first place by this order of mixing. The final products from each experiment were compared and found to be identical in all respects, dissolving in alcoholic potash with the orange-red colour intermediate between the orange solution of the dimetadinitro-compound and the magenta-red solution of the diparadinitro-compound in the same solvent. This pair of experiments serves to show that the para-metadinitro-compound is a distinct chemical substance having a melting point of 212-213°.*

Experiment IV.-The next experiment is a repetition of a former attempt (this Journ., Trans., 1887, 440, note) to synthesise the parameta-compound by mixing the two normal isomerides

(p)NO2•C ̧H ̧•N2H·C ̧H ̧·NO2(p) and (m)NO2·CH, N2H·C ̧H ̧·NO2(m). 2 grams of each were completely dissolved in alcoholic soda, the solution diluted with water, and then just neutralised with acetic acid so * In our former papers we gave the melting point as 211°.