8 i 9 LXXIV.-The Gonstitutiola of the Terpmes and of Benzene. By WILLTAX A. TILDEN D.Sc. F.R.S. THE idea that turpentine and lemon oils consist of a hydride OE cymene is due to Oppenheim (Ber. 5 94 and 628). This view has been supported by KekulB (Bey. 6 43&) and very generally adopted down to the present time. Nevertheless I believe it to be incon-sistent with many of the most prominent characteristics of these hydrocarbons and so long ago as FebruarT 1878 I laid before the Chemical Society a statement of some of the difficulties which beset this hypothesis. The formulae which I then proposed for the terpenes I do not now pretend to sustain inasmuch as they involve the mistaken assumption that these compounds are sexvalent cr contain three double linkings of carbon.Nevertheless my opinion is not only unchanged but is fortified by all that I have since learnt of these hydrocarbons that whilst the terpeness almost certainly contain a nucleus of six atoms of carbon these are not disposed in the manner in which they ai-e assumed to be in benzene whatever constitutional formula is adopted for that substance. The publication of a number of papers on the chemistry of the terpenes more especially the very important researches of Wallach (Annulen 225 227 230 238 239 241 245 246) hasattracted my attention again to the subject and I desire to state in the first instance the grounds for my belief that the terpenes are not consti-tuted as benzene derivatives. The hypothesis commonly received originated in the experiments of Oppenheim Barbier and KekulB upon the production of paracymene from turpentine.C. R. A. Wright afterwards examined the action of bromine upon a number of terpenes from various sources and found that they all gave the same cymene in different proportions. Nevertheless the importance of this fact must not be exaggerated, for it must be remembered that the yield in the most favourable case is much below the theoretical arid cymene cannot be reconverted by any known process into a terpene. Moreover the known hydrides of benzene toluene &c. are totally unlike terpenes in character. The hypothesis also accounts only for three dihydrides assuming the propyl-group to remain the same or six if there be a cymene con-* Using this term to include all the liquid members of the group natural 01.artificial but not the solid “ camphenes,” which hare lotally different characters. I shall generally use Wallach’s nonienclature 880 TILDEN THE CONSTlTUTION OF taining isopropyl. The isomeric terpenes already recognised exceed this number. A second point upon which great stress has been laid is the nature of the oxidation products of the terpeiies and their derivatives terpin hydrate &c. As the impression seems to preva>il that the terpenes when acted upon by dilute nitric acid form a productive source of para-toluic acid I have made a few comparative experiments with the object of determining the proportion of this acid obtainable under exactly the same conditions from several terpenes from cymene and from paraxylene. A weighed quantity of the hydrocarbon was in each case boiled in a flask with reflux condenser with a quantity of nitric acid slightly in excess of that theoretically required for its oxidation to toluic acid.The nit& acid was diluted with four times its volume of water. When the hydrocarbon had disappeared the mixture was allowed to cool the crystailised acid filtered off dissolved in boiling dilute solution of ammonia then reprecipitated by dilute nitric acid, t+nd after some hours collected dried and weighed. The acid thuR obtained was white ; it melted a t about lXO and was therefore nearly pure. Quantities of hydrocarbon varying from 15.9 to 17.1 grams were taken. The following proportions are calculated from the results :-100 parts (by weight) give of paratoluic acid-Cyrnene (fiwm turpentine) .80.1 Cymene (from cummin oil Kahlbaum) 73.1 Psraxylene (Kahlbaum's pure) . 77.2 Australene (from American turpentine). . Terebenthene (from French turpentine). . 1.2 1-9 0 Hesperidene b. p. 175" (from orange oil) . . The hesperidene having failed to give any recognisable quantity of toluic acid an experiment on a larger scale was tried. About 85 grams of the hydrocarbon gave by similar treatment an acid liquid which when concentrated to a smdl bulk deposited 0.15 gram of an insoluble acid (probably terephthalic acid). The chief products of oxidation were a yellow resin which by protoracted action of strong nitric acid gave a viscid nitro-compound and oxalic acid (7.07 grams of anhydrous H,C,OJ) and a small quantity of terebic acid.A comparison was also made between the products of oxidation by chromic acid solution upon cymene and australene with the following result :-Cymene 17 grams gave 7 grams ferephthalic acid or 41 per cent THE TERPENES AND OF BENZENE. 881 Australene 16.3 grams gave a trace only insufficient to weigh, together with a minute quantity of a camphor in the conderiser. Wright (Trans. 1873 553) in oxidising pure hesporidene by chromic acid mixture obtained no terephthalic acid acetic acid being the only product. My contention has always been that the natural terpenes do not yield appreciable quantities of aromatic acids by oxidation ; that they do yield by isomeric change under the influence of heat o r of reagents, compounds which must be regarded either as direct derivatives of benzene or closely connected with it must of course be admitted.Thus not only is cymene obtainable by the removal of 2 atomfi of hydrogen but t,he dipentene (terpilene) obtained from the dihydro-chloride got from turpentine doubtless contains a nucleus of six carbon atoms which either has assumed partly the structure of benzene or possesses a stmwture which is easily transformed into that of benzene. I found that when oxidised by nitric acid as already described, 100 parts of it gave 27-6 of toluic acid although the turpentine from which it was derived gave practically none. The characteristics of the terpenes as a class may be stated as follows. The terpenes are polymerised very readily by the action of heat or of small quantities of sulphuric acid and other agents.All the terpenes combine instantly and eagerly with chlorine and bromine whilst benzene and cymene behave save under exceptional circumstances as saturated compounds. All the terpenes unite with hydrochloric hydrobromic and hydr-iodic acids. Some of the terpenes (the pinene-group especially) also unite slowly with the elements of water forming two alcoholic bodies terpineol, C,,H,,OH and terpin CloH,,( OH),. All these characters are precisely those which are exhibited so plainly by the olefines but in no case by hydrocarbons directly derived from benzene. A few other facts deserve notice. No natural terpene is recoverable from its compound with bromine or hydrochloric acid. A new hydrocarbon of the same composition, Cl0HI6 is always formed namely a camphene dipentene terpinene, or terpinolene according t o circumstances.Again the terpenes are resolved when heated to a temperature of 400" to 500' into a pentene, C5Hs which may be retransformed into a dipentene Cl0H16 the relation between these two compounds being exactly similar to the relationship subsisting %etween amylene and diamylene. If we recall the existence of the sesquiterpenes and the constituents of colophene which are polymerides of terpene the persistence of the YOJI. LIII. 3 882 TlLDEN THK COSSTITUTION O F factor 5 in the number of carbon atoms C5H8 C,,H:,, CI5HZ4 C,,H,?, &c. is remarkable. The combining capacity of the majority of the terpenes appears t o be pretty definitely indicated by the formation of compounds with 2B C1 and in four cases namely limonene (citrene) dipentene sylves-trene and terpinolene with Br4.These tetmbromides which were d l discovered by Wnllzch are crystalline and stable. Nevertheless, it occurred to me as just possible that these hydrocarbons might after all possess a greater combining capacity though the products of their combination with bromine niiglit not be stable enough to admit of their being isolated. A new estimation of their capacity f o r bromine seemed also very desirable in view of the conclusion arrived a t by Gladstone (Trans. 1886 p. 609) from the refraction equivalent and recently restated by Briihl ( B e y . 21 156) to the effect that the pinenes (australene terebentllene &c.)? contain only one double bond, or in other words are only bivalent.T o determine the capacity for bromine a solution of hypobromite of sodium was used. The pure hydrocarbon diluted with twice its volume of chloroform mas mixed with diluted hydrochloric acid in a bottle and the hypohromite solution added in small qiiantities shaking after each addition till the hydrocarbon which settled clown exhibited a faint yellowish colour. This plan being not quite satisfactory it was found better to add at once a slight excess of the hypobromite 80111-tion shake well f o r a minute and then add some potassium iodide, and estimate by thiosulphate solution the amount of liberated iudine. Four to five grams of the hydrocarbon were taken for each experiment. By this method 136 parts or 1 molecule of hydrocarbon took in each case the following quantities of bromine :-Australene (from American turpentine) Terebenthene (from French turpentine) Hespericlene (from orange oil) Dipentene (from dihydrochloride).. 297 77 Br4 EO x 4 320 313 307 326 Calculated for Br 80 x 2 . . . . . . . . . . . . 160 There can be no doubt in t'hese cases that the hydrocarbon fixes four atoms of bromine and no more. One molecule or 154 parts of terpineol C,,H,,OH combined with 185.6 parts of bromine which is only a little more than is required on the assumption that it unitfes with Br2 or 160 parts. An aqueous solntion of tcrpin CI,H,,(OH), does not decolorise bromine-watey even after some t,inie. We have therefore Quadriralent. Bivalent. Saturated THE TERPENES AND OF BENZEKE.883 So far the terpenes. It is well known however that there exists another class of compounds the camphenes CloH,, which are iso-meric with the terpenes and derivable from them. It is unfortu-nate that these compounds should be so commonly classed with the t,erpenes and apparently regarded by many writers as pmsessing a, simiiar structure for t'heir chemical characteristics are very different. When American o r French turpentine oil free from water is satu-rated with dry hydrochloric acid gas a solid compound is formed con-taining CI,H,,HC1. From this the elements of hydrochloric acid may be withdrawn though with difliculty by the action of alco-holic potash or of sodium acetate &c. The resulting hydrocarbon is camphene optically active or inactive according to circumstances.It is a solid melting a t 48" arid boiling a t 160". The same substance is formed in considerable quantity among the products of the action of sfrong sulphuric acid upon turpentine (Arnistrorq arid TiIden Trans. 1879 7 3 3 ) . It is also formed by the ackion of sodium upon the hydrochloride C,,,H,,HCI from turpentine aiid i s obtainable through the chloride from borneol and hence from camphor. Camphene differs notably from the terpenes in its chief proauct of oxidation for it yields neither toluic nor terephthalic acid but acetic acid and camphor. It does not combine with bromine but forms an ill-defined liquid monobromo-compound C,,H,,Br (Wallach, Ann. 230 235). It does unite with 1 mol. of hTdrochloric acid forming a com-pound which is very similar in appearance to the monohydrochlo-ride CIOH:IGHCI prepared from turpentine oil but differs from that compound in dissociating into a hydrocarbon and HC1 when heated' (Ehrhardt Chem.Netos 54 239) and in. yielding up readily the elements of HCl when boiled with water or alkaline solutions regene-rating the camphene. There are three varieties of this hydrocarbon which a p e e com-pletely in all physical and chemical characters except in their action on polarised light the one possessing right-handed rotatory power, another being left-handed whilst the third is inactive. When turpentine unites with dry hydrochloric acid therefore a change is induced manifestly more profound than that which occurs when it combiiies with 2HC1 as it does in the presence of water or alcohol.This latter compound when heated alone or with aniline readily yields liqiiid dipentene which in general characters is veiy similar to the original hydrocarbon and is probably identical with the optically inactive product of the action of heat upon turpentine. We must therefore recognise the following chemically distinct isomerides (see also Wallach Ann. 239 45). 3 r 884 TILDEN THE CONSTITUTIOK OF I. NATURAL TERPENES (optically active liquids). 1. Pinene (australene and terebenthene). 2. Limonene (citrene). 3. Sylvestrene (from Swedish turpentine). 4. Phellandrene (from Phellandriurn aquaticunz &c. Pesci, Gazz. Claim. 16 225). 11. ARTIFICIAL TERPENES (optically inactive liquids). 1.Djpentene (from the dihydrochloride CloHf1162HC1 m. p. 2. Terpinolene (from the action of sulphuric acid on terpin). 48'). 3. Terpinene ( 7 7 9 7 9 ) 1. ITI. ARTrFrCrAL CBMPHENES (solids-two optically active in opposite directions-one inactive). The combining capacity of none of these exceeds four units. Con-sequently it is not possible to represent the nucleus of six carbon-atoms which undoubtedly they all contain as forming an open chain. They must be united into a closed chain containing a t the most two double bonds. Hence we are driven to such formule as those which have been proposed by Oppenheim (Bsr. 5 98) by Goldschmidt and Z:ii&er (Ber. 18 1729 and 20 486) by Wallach (Annalen 239 46), and more recently by Briihl (Bey. 21 165). Goldschniidt and Zurrer derive the formula of lirnonene (= carvene or citrene) from that of carvol in consequence of the relations which they have cstab-lislied between these t w o compounds.Thus :-Limonene. Carvol. QsH7 c3137 c c' C H / ~ H CH+H cH\ c IT CH+ HC*CH3 HC.CLI, Unfortunately the constitution of carvol as a benzene-derivative is not fully established for its conversion into carvacrol is attended by :a very energetic reaction in the course of which we may well believe that a profound alteration of structure occurs. Wallach has shown in a very plausible way how the polymerisation of isoprene C,H (assumed to be CH,:C(CH,)*CH:CH,) into dipen-tene may be accounted for by a similar formula. I do not agree with him in some points of detail for example in his statement that pinenes contain only one double bond for as shown on R previous page the pinenes combine with the same amount of bromine as the citrenes and they form with hjdrochloric acid the same dihy THE TERPENES ASD OF BENZESE.885 drochloride. Neither can I accept his formula for solid camphene, which certainly contains no double bond at all as testitied by its i n -difference to bromine. But I am glad to observe that he adopts the view which I have so long maintained that pinene (turpentine oil " keiu gewohnliches Hydrocymol ist " (Ann. 239 48). The formulae for the terpenes range themselves under sevcral t,ypes, which represent them all as containing two double bonds such as are conveniently called ethylenic bonds. The most probable of these are represented as follows in which R and R' stand for the t w o groups methyl and propyl.Of course okhers may be conceived as for example by exchange of position by R and R' in 1. A t present however we do not know whether all the different species have yet been recognised. There may be other isotnerides which must hereafter be taken into account and there is very lithle to guide us in assigning to each known hydrocarbon its appropriate formula. It will be sufficient therefore if I indicate the manner in which they may be used by rcference to one example. Adopting a formula of type I for turpentine-and indicating the asymmetric atom of carbon which it contains by ,z black-faced letter its transformation into the monohydrochloride and t h i s subsequently into camphene and campbor may be represented by the following formule : 886 TILDEN THE COXSTITUTIOS OF But it will naturally be remarked that such formule as these, which have been proposed for the terpenes are the accepted formuh for derivatives of benzene.Oppenheim’s formula ‘or turpentine which is identical with that given above except that the positions of the methyl and propyl are reversed was deliberately contrived for the purpose of exhibiting the relation of turpentine to barizene as indicated by its conversion into cymene. The conclusion t o which we are impelled is that Kekul6’s formula for benzene must be abandoned. If the evidence for this con-clusion rested solely upon our present imperfect knowledge of the terpenes it would of course possess but little value.This however is not the case. The chief objection which has been raised against Reknl6’s formula, namely that the positions 1 2 and 1 6 are not identical has been met by a well-known hypothesis due to Kekul6 himself. That hypothesis has very little probability in its favour but whether that explanation is now generally accepted is not of great importance in the presence of anot’her objection which has not been so generally recognised as I think it deserves to be. KekulB’s formula represents benzene as containing three times over the same kind of linking of the carbon atoms which we must admit in the case of ethylene, VIZ. :- -c=c-. It is characteristic of this grouping that it is associated with the capacity for instantaneous combination with bromine with tjhe liydracids &c.In benzene we have a compound which exhibits no trace of this character. Bromine dissolves in benzene without form-ing addition products after the mariner of the olefines and other hydrocarbons in which we assume the existence of double or treble linking and only under exceptional circumstances are these addition products formed. In the formuk of many aromatic compounds the same symbol is used to represent two distinct functions. Thus in cinnamic acid for example THE TERPESES AND OF THE BESZENES. 887 CH CH/\C-CH= CH -CO~H CAI ICH ‘.c//H the atoms of carbon united by two bonds in the side-chain combixe instantly with HBr or HI with Br2 or with HC10. They are in fact strictly et’hylenic in character. The two double bonds in the hexagon correspond to nothing of the kind for they do not unite in the same way under similar circumstances.Doubtless most chemists have been in the habit of indulging in a mental reservation as to the significance of the double bond in a closed chain. This however, does not justify the employment of the same symbol to express two different things. Again when benzene and its homologues are submitted to oxidation, the six carbon atoms remain very persistently united and a series of acids containing a nucleus of Cs results. I n other cases where double linking unquestionably occurs i t seems very often to be a point of weakness the carbon atoms thus united separating more readily than those which are united by single bonds. Ally1 alcohol for example, gives by the action of dilute nitric acid formic (or carbonic) and oxalic acids.Now this splitting up of the double bonds in a forniuln like that which Kekulh attributed to benzene ought to yield abund-ance of oxalic acid. This is what the terpenes do but which benzene and its homologues do not. v. Baejer in some important papers (Bey. 19 1797 and more recently Anrcalen 245 103) describes the remarkable properties of the hydroterepht,halic acids and points out that these compounds no loriger possess the character of benzene derivatives but in spite of the ring-like arrangement of the carbon atoms their chemical nature indicates that they must be regarded as compounds of the fatty series. There is in fact the same sort of relation between the properties of terephthslic acid and dihydroterephthalic acid that there is bet ween cymene and the terpenes.I n order to explain the peculiarities of the hydroterephthalic acids, v. Rae-yer proposes a theory which seems to be nearly identical with that enunciated by Armstrong about a year previously (Phil. Mag., Feb. 1887). I n benzene and its immediate derivatives he assumes that the six carbon atoms form a symmetrical figure. One unit of valency of each draws the carbon atoms towards the centre of the ring but v. Baeyer expressly states that they must not be assumed to be free in the ordinary sense neither are opposite carbons united together the central unit’s of valency are only to be regarded as passive. It seems to me as already pointed out by Ladenbur 888 TILDEN THE CONSTITUTION OF THE TERPENES ETC.(AnnuZen 246 382) very difficult if not impossible to distinguish this formula from the diagonal formula of Claus. I do not tliink tlie present state of knowledge in regard to this matter enables us to state more than the following propositions :-1. The six atoms of carbon in benzene form a very stable group, which is not easily broken up by heat or by chemical agents. 2. The carbon atoms are all in the same condition and the hydrogen is distributed equally among them. 3. The carbon atoms are not united together by ethylenic bonds. 4. There are three and only three disubstitution-derivatives. Either of the formulae of Claus of v. Baeyer of Armstrong or the prism of Ladenburg satisfies these conditions but a t present there seems no solid reason for preferring one before the other.With regard to the Claus formula v. Baeyer has already pointed out that in any ring-like arrangement the carbons in the pya-position must be further apart than those in the oi-tho-position and hence the existence of three isomeric diderivatives may be accounted for. Briihl’s two recent papers on the terpenes and their derivatives (Bey. 21) contain suggestions as to the constitution of the terpenes, based upon the observation of their refraction equivalents. These p,ipers have been already very severely criticised by W allach (Annulen, 245 121) and I need only add that I agree with the latter in thinking that much reliance cannot be placed upon the recorded values of the refraction equivalents in this group of compounds by reason of the more than doubtful puritly of many of the substances operated upon.And when the conclusions are in direct conflict with those derived from the most obvious chemical properties as for example when the saturating power of camphene which is untouched by bromine is represented as being the same as that of terebenthene, which instantly combines with four atoms of bromine all we can do is to wait for further information as to their optical properties and the relation in which these stand to chemical constitution. Evidence from other sources so far as it is available supports the view that I have adopted. Thus Hartley (Proc. R o y . Soc. 1879, 29 290) found that australene terebenthene and hesperidene gave none of the bands which are exhibited in the spectrum of tlie ulf ra-violet by cymene and other undoubted benzenoid compounds. And from Abney and Festing’s (Phil. Trans. 1881) observations at the opposite extremity of the spectrum the crucial line 867 with its attendant band characteristic of benzene-derivatives is not seen in the absorption-spectrum of turpentine