September 14, 1893J 



NA TURE 



479 



posal with a short account of a line of study in what I have 



already termed "comparative chemistry," which is not only of 



! inherent interest, but seems to give us the means of filling in 



1 some details of a hitherto rather neglected chapter in the early 



i chemical history of this earth. 



I The most remarkable outcome of " comparative chemistry " is 

 \ the periodic law of the elements, which asserts that the properties 

 I of the elements are connected in the form of a periodic function 

 j with the masses of their atoms. Concurrently with the 

 I recognition of this principle, other investigations have been in 

 I progress, aiming at more exact definitions of the characters of 

 I the relations of the elements, and ultimately of their respective 

 I offices in nature. Among inquiries of this kind the comparative 

 study of the elements carbon ani silicon appears to me to 

 i possess the highest interest. Carbon, whether combined with 

 I hydrogen, oxygen, or nitrogen, or with all three, is the great 

 I element of organic nature, while silicon, in union with oxygen 

 I and various metals, not only forms about one-third of the solid 

 crust of the earth, but is unquestionably the most important ele- 

 ment of inorganic nature. The chief functions of carbon are 

 those which are pe-formed at comparatively low temperatures ; 

 hence carbon is essentially the element of the present epoch. 

 i On the other hand, the activi'ies of silicon are most marked at 

 i very high temperatures ; he ice it is the element whose chief 

 ' work in njture was performed in the distant past, when the 

 temperature of this earlh was far beyond that at which the 

 carbon compounds of organic life could exist. Yet between 

 I these dominant elements of widely different epochs remarkably 

 I close analogies are traceable, and the characteristic difference-i 

 i observed in their relatione with other elements are jast those 

 I which enable each to play its part effectively un ler the con- 

 ditions which promote its greatest activity. 



The chemical analogies of the tw i tetrail elements carbon and 

 silicon are most easily recognised in compounds which either do 

 not contain oxygen, or which are oxygen compounds of a very 

 simple order, and the following table will recall a few of the 

 most important of these, as well as some which have resulted 

 from the fine researches of Friedel, Crafts, and Ladenburg: — 



Some Silicon Analogues of Carbon Compounds. 

 SiHj Hydrides CHj 



i:a :;: :;:} chlorides { •;; ca, 



S1O2 Oxides CO2 



WjSiOj Mela Acids . H^COj 



nSiHO, Formic Acids ... HCHOj 



(SiHO)„0 Formic Anhydrides (CHO).p? 



HjSijOj Oxalic Acids ... HjCjOj 



HSi(CH3)Oj ... Acetic Acids .. HCfCHJO, 



HSi(CeH5)0., ... Benzoic Acids ... HC(C6H,)0, 



SiCsH^H Nonyl Hydrides ... C.h'h 



SiCgHigOH ... Nonyl Alcohols ... CuHijOH 



But these silicon analogues of carbon compounds are, 

 generally, very different from the latter in reactive power, 



;whereas the analogous marsh gas does not take fire in air 

 |below a red heat. Again, the chlorides of silicon are rapidly 

 .attacked by water affording silicon hydroxides and hydrochloric 

 ,acid ; but the analogous carbon chlorides are little affected by 

 .water even at comparatively high temperatures. Similarly, 

 5iliconchloroform and water quickly produce silico-formic acid 

 'and anhydride along with hydrochloric acid, while ordinary 

 Ichloroform can be kept in contact with water for a considerable 

 time without material change. 



Until recently no well-defined compounds of silicon were 

 known including nitrogen ; but we are now acquainted with a 

 number of significant .substances of this class. 



Chemists have long been familiar with the fact that a violent 

 reaction takes place when silicon chloride and ammonia are 

 illowed to interact. Persoz, in 1830, assumed that the resulting 

 white powder was an addition compound, and assigned to 

 !t the formula SiCl^, 6 NI{„ while Besson, as lately as 1892, 

 'ave SiCl^, 5 NHj. These formulae only express the proportions 

 jn which ammonia reacts with the chloride under different 

 ,:onditions and give us no information as to the real nature 

 >f the product; hence they are al nost useless Other 

 rhemists have, however, carefully examined the product 



N^. I 246 VOL. 48] 



of this reaction, but owing to peculiar difficulties in the 

 way have not obtained results of a very conclusive kind. 

 It is known that the product when strongly heated in a current 

 of ammonia gas affords ammonium chloride, which volatilises, 

 and a residue, to which Schutzenberger and Colson have as- 

 signed the formula SijNjH. This body they regard as a definite 

 hydride of SijNj, which latter they produced by acting on 

 silicon at a white heat with pure nitrogen. Gattermann suggests 

 that a nearer approach to the silicon analogue of cyanogen, 

 Si2N2, should be obtained from the product of the action of 

 ammonia on silicon-chloroform ; but it does not appear that this 

 suggestion has yet borne fruit. It was scarcely probable that 

 the above mentioned rather indefinite compounds of silicon with 

 nitrogen were the only ones of the class obtainable, since bodies 

 including carbon combined with nitrogen are not only numerous 

 but are among the most important carbon compounds known. 

 Further investigation was therefore necessary in the interests of 

 comparative chemistry, and for special reasons which will appear 

 later on ; but it was evident that a new point of attack must be 

 found. 



A preliminary experimental survey proved the possibility of 

 forming numerous compounds of silicon containing nitrogen, 

 and enabled me to select tho e which seemed most likely to 

 afford definite information. For much of this kind of work 

 silicon chloride was rather too energetic, hence I had a con- 

 siderable quantity of the more manageable silicon lelrabromide 

 prepared by Serullas' method, viz. by passing the vapour of 

 crude bromine (containing a little chlorine) over a strongly 

 heated mixture of silica and charcoal. In purifying this pro- 

 duct I obtained incidentally the chloro-bromide of silicon, 

 SiClB'3, which was required in order to complete the series of 

 possible chlorobromides of silicon.^ 



Silicon bromide was foun 1 to pr')duce addition couipounds 

 very readily with many fee .|y basic substances containing nitro- 

 gen. But one group of bromides of this class has yet been 

 investigated in detail, namely, the pro lucts afforded by thioureas. 

 The typical member of this groap is the perfectly definite but 

 uncrystalline substance 



SiBr /(CSN2H^)4Br 



Substituted thioureas afford similar bodies, the most interest- 

 ing of which is the allyl compound. This is a singularly viscid 

 liquid, which requires several days at ordinary temperatures to 

 regain its level, when a lube containing it is inverted. But 

 these are essentially addition compounds, and are therefore com- 

 paratively unimportan*. 



In most cases, however, the silicon haloids enter into very 

 definite reaction with nitrogen compounds, especially when the 

 latter are distinctly basic, such as aniline or any of its homo- 

 logues. One of the principal products of this class of change 

 is the beautiful typical substance on the table, which is the first 

 well-defined crystalline compound obtained in which silicon is 

 exclusively combined with nitrogen. Its composition is 

 Si(NHC8H5)4." Analogous compounds have been formed with 

 the toluidines, naphthylamines, &c., and have been examined in 

 considerable detail, but it suffices to mention them and proceed 

 to point out the nature of the changes we can effect by the 

 action of heat on the comparatively simple anilide. 



When silicon anilide is heated carefully in vacuo it loses one 

 molecule of aniline very easily and leaves triphenylguanidine, 

 probably the a modification ; if the action of heat be continued, 

 but at ordinary pressure and in a current of dry hydrogen, 

 another molecule of aniline can be expelled, and, just before 

 the last trace of the latter is removed, the previously 

 liquid substance solidifies and affords a silicon analogue of the 

 insoluble modification of carbodiphenyldiimide, which may 

 then be heated moderately without undergoing further material 

 change. A comparison of the formulce will make the relations 

 of the products clear : — 



Silicotetraphenylamide— Si(NHPh)4 

 Silicotriphenylguanidine — Si : NPh. (NHPh)j 

 Silicodiphenylaiimide — Si : (NPh)j. 



Moreover, thediimide has been heated to full redness in a gas 

 combustion furnace while dry hydrogen was still passed over it ; 

 even under these conditions little charring occurred, but some 



■ Three years later Besson formed th^ same compound, and described it 

 as new. 



2 Harden has obtained an uncrystalliuc intermediate compound, 

 SiC'a'NHCeHflVj. 



