solutions thus obtained in many ways resemble hypoiodite solutions in their behaviour towards arsenite, and towards potassium iodide and sodium hydrogen carbonate, only they are far more stable. With ammonia, the formation of iodate and iodide is very slow, the change not being complete at the end of two weeks. With methylamine, the power of oxidising arsenite slowly diminishes, but in three weeks has not entirely disappeared. Iodate is only formed in very small quantity, and cannot be detected for the first few days. Iodide is produced in the solution from the product formed initially.

(1) 3N aqueous ammonia. Temperature 12-14°. Total iodine per 100 c.c.= 0.9 c.c. of N/10 iodine

=

(0.0114 gram).

Per cent. of iodine as iodate and iodide.

0.0

13.3

22.2

35.5

93.3

These titrations were carried out with 300 c.c. of the solution, using N/100 standard solutions.

(2) 3M methylamine solution. Temperature 15°.

Total iodine per 100 c.c. = 204 c.c of N/10 iodine (=0·259 gram).

It was thought that possibly in the case of methylamine the arsenite was oxidised by iodate, but this was shown not to be the case by a blank experiment. Also the same amount of oxidation of arsenite took place when excess of concentrated sodium hydroxide was added to the methylamine solution before the arsenite was run in. It appears then that the initially formed product, whether methyldiiodoamine, CH, NI,, or hypoiodite, is fairly stable, and is only slowly converted into hydriodic acid (methylamine hydriodide) probably in oxidising some of the methylamine. This point will be further investigated.

Reaction of Iodine with Mercuric Oxide.

Köne (Poggendorff's Annalen, 1845, 66, 302) and Lippmann (Compt. rend., 1866, 63, 968) have noticed the formation of hypoiodous acid (or mercury hypoiodite) when iodine is shaken up with water and mercuric oxide. In the presence of amylene, Lippmann observed the formation of an iodohydrin. We have estimated the amount of hypoiodite and iodate in solutions prepared by shaking up finely-powdered iodine and mercuric oxide (precipitated) with water, alcohol being omitted owing to the action of hypoiodite on it. The shaking was continued for a few minutes, and the solution then filtered through asbestos or hardened filter paper. The solutions were neutral or faintly acid, the latter being the case when the shaking was not prolonged. They were colourless and gave no reaction with starch paste on acidifying or on adding sodium hydrogen carbonate. For some hours (3—4), they gave the reaction for hypoiodite with sodium hydrogen carbonate, starch paste, and potassium iodide. On addition of potassium iodide to the acidified solution, it became deep brown from the liberation of iodine, a reaction which points to the presence of iodate.

A solution prepared by shaking iodine with mercuric oxide for 5 minutes contained iodine per 100 c.c. 2-3 c.c. of N/10 iodine, of which 48 per cent. was present as hypoiodite, and the remainder as iodate. In a similar experiment, after 15 minutes shaking, 4·4 c.c. of N/10 iodine were contained in 100 c.c., 13-14 per cent. of which was in the form of hypoiodite. The filtering of the solution, &c., occupied some 10 minutes after the cessation of the shaking.

From these results, it would seem that the solutions obtained from iodine and mercuric oxide contain only a small quantity of hypoiodite, and that the iodine is chiefly present as iodate. Iodide is only present in small quantity, owing to the low solubility of mercuric iodide.

No attempt was made to determine the velocity constant of these transformations of hypoiodite. Schwicker (Zeit. physikal. Chem., 1895, 16, 303) has attempted to determine this constant for the transformation of potassium hypoiodite, and found that it varied greatly with the variations in initial concentration of the hypoiodite and the alkali. Excess of potassium iodide was present in the solutions. It is possible that the investigations of the solutions of calcium hypoiodite or ammonium hypoiodite (?) which undergoes transformation relatively slowly may lead to some elucidation of this point.

CHEMICAL LABORATORY,

ST. BARTHOLOMEW'S HOSPITAL AND COLLEGE, E.C.

LXXVI. Researches on Silicon Compounds. Part VI. On Silicodiphenyldiimide and Silicotriphenylguanidine. By J. EMERSON REYNOLDS, D.Sc., M.D., F.R.S.

IN Part V of this series of papers (Trans., 1889, 55, 474), I described the well defined, crystalline silicophenylamide, Si(NH CH2), which proved to be the first of a new class of compounds in which silicon is exclusively united with nitrogen. In the course of the paper on the phenylamide, I pointed out that when the crystals were heated somewhat beyond their melting point, 137°, they easily lost some aniline, and added that "the residue consisted of a mixture of silicon compounds to be described later on." The nature of each of these products was, in fact, afterwards made known in the course of my Presidential Address to Section B of the British Association at Nottingham, but no details were then given, and I now desire to repair that omission preliminary to laying before the Society an account of more recent investigations.

The primary interest which attaches to the examination of the changes effected by heating the silicophenylamide is due to the probable production of silicon analogues of cyanogen compounds.

In the preliminary experiments, it was observed that about onefourth of the total nitrogen of the anilide was easily driven off by heat in the form of aniline, and that a rather higher temperature was necessary for the removal of another fourth. At the end of this second stage, the previously fluid mass suddenly solidified, but continued heating of this residue led to the expulsion of more aniline. Of the three stages of decomposition recognised, that of solidification was most strongly marked, and therefore was first examined in detail.

Twelve grams of silicophenylamide, Si(NH-CH5)4, were heated in a platinum boat placed in a glass tube through which a current of carefully dried hydrogen slowly passed. The gas carried with it the aniline evolved on decomposition, and the latter was condensed in a U-tube, whence it fell into a small measuring vessel. The process was stopped when the aniline collected represented half the nitrogen present in the weight of the anilide taken, and this point was found to coincide with that of the solidification of the residue already mentioned. The aniline collected contained a small quantity of a silicon compound, carried over in the current of gas and vapour.

The residue when cold was broken up and digested with carefully dried benzene, which dissolved little of the material, but removed adherent aniline, unchanged anilide, and another substance. When

purified by this treatment, the product was a white, amorphous powder, quite insoluble in benzene, light petroleum, carbon disulphide, ether, or alcohol. When heated with alcoholic solution of sodium or potassium hydroxide, it gradually dissolved with separation of aniline and evolution of a slight odour of phenyl isocyanide.

0.2912 gave 0.0816 SiO2. Si = 13.12.

0.351 0.263

0-0993 SiO2. Si=13.2.

30-8 c.c. moist nitrogen at 15° and 767 mm. N=13·76. SiN2(CH)2 requires Si = 13.33; N = 13.33 per cent.

6

The nitrogen is rather high, but this was subsequently explained by the presence of traces of another compound richer in nitrogen, which is formed at an earlier stage of the decomposition, but was unknown at the time the analyses were made, and was not completely removed by benzene.

The new substance may obviously be regarded as the silicon analogue, either of diphenylcyanamide or of carbodiphenyldiimide.

The genesis of the compound from monophenylic groups and the separation of aniline, rather than diphenylamine and ammonia, on decomposition by alkalis, lead to the conclusion that it is silicodiphenyldiimide. Hence the action of heat under the conditions stated is essentially represented by the equations :

Si(NH•C,H,)4

=

Si(N•CH)2 + 2CHNH,.

5/2

Another experiment, in which the heating of the phenylamide was stopped just short of the solidification point, gave a residue which was partially soluble in benzene. The latter solution, when evaporated, left a residue similar in appearance to the diimide and, like it, contained 13.1 per cent. of silicon, but it could not be again dissolved in benzene. Hence, an unstable soluble modification of the diimide appears to exist which easily changes into the compound which is insoluble in benzene.

As the heated amide more easily lost the first molecule of aniline than the second, it seemed probable that the following change could be realised and silicotriphenylguanidine or an isomeride be obtained.

Si(NH•C,H,)=C,H •N:Si(NH-CH,),+CH•NH,.

6

Eight grams of the pure silicophenylamide were carefully heated as before in a current of dry hydrogen. When rather more than one molecular proportion of aniline had been collected, the decomposition was stopped, and the residue allowed to cool in hydrogen. There was

no thickening of the residue while hot, but it solidified to a yellowish, transparent mass on cooling.

The product was broken up and heated with benzene. This left a small portion undissolved which possessed the characters of the diimide. The benzene solution, when mixed with twice its volume of light petroleum, gave a somewhat crystalline precipitate totally unlike the anilide obtained under similar conditions. This substance was washed with light petroleum to remove any free aniline, again dissolved in benzene, and four volumes of light petroleum were then gradually added. A precipitate formed, which was seen under the microscope to consist chiefly of droplets of a somewhat sticky substance. On standing for some hours, the liquid afforded a number of small crystals, which separated on the sides of the flask in which the precipitation was effected. These crystals were mechanically removed and purified by dissolution in benzene and subsequent separation on addition of light petroleum. This treatment involved much loss owing to the very sensible solubility of the compound in light petroleum.

The purified crystals were short prisms quite distinct in appearance from the original silicophenylamide, and much less completely precipitated from their benzene solution by addition of light petroleum. It did not afford any definite compound with acids, but was decomposed to some extent even by dry hydrogen chloride, as well as by ordinary solvents.

The crystals melted at 230°, and commenced to decompose at a slightly higher temperature; the residue did not solidify until cooled to 131°.

On analysis, the purest crystals gave Si = 9·11 and N = 14.03 per cent. CH, N:Si(NH.CH)2 requires Si = 9.24; N = 13.86 per cent. The substance is therefore a silicon analogue of a triphenylguanidine in nearly pure condition.

Hence the carefully regulated decomposition of the silicophenylamide by heat leads to the formation of

Silicotriphenylguanidine ... CH·N:Si(NH·C2H5), and

Silicodiphenyldiimide

Si(N⚫C&H)2.

When the action of heat was studied under diminished pressure, at 16 mm., the guanidine stage was quickly passed, and the second molecular proportion of aniline carried over with it a thick liquid of high boiling point, before the residue suddenly changed to the solid diimide. This liquid was easily freed from aniline by careful heating at the low pressure maintained, and when cold was still a viscid substance. This proved to be soluble in benzene, but was precipitated in oily droplets by light petroleum, and adhered to the sides of the vessel in which the substance was thrown down. These droplets solidified after some hours, and then no longer dissolved in benzene.