ELIMINATION OF THE CARBETHOXYL QROUP E"0. 143 XV1.-Experiments on the Elimination of the Carb-ethoxyl Group from Tautomeric Systems. Part I. Derivatives of Inden e. By CHRISTOPHER KELE INGOLD and JOCELYN FIELD THORPE. THERE have been placed on record within recent years (T. 1905, 87 1669 1685; 1911 99 2187 and subsequent papers of the same series) a number of experiments dealing with substances possessing the kind of tautomerism which is associated with the three-carbon system CHCC. These experiments have for the mmt part dealt with glutaconic acid and its alkyl derivatives and the conclusion was reached that glutaconic acid itself has the symmetrical or " normal " structure (11) t'he unsaturated or " labile " form (I) being too unstable t o have any but a momentary existence.CO,H*CWCH:CH-CO,H CO,H b~ CH,. t ! ~ CO,H . (I. 1 (114 When however alkyl groups were introduced into the three-carbon system the unsaturated form was found to become very noticeably more stable with increasing weight of the substituents. It was thought desirable to extend these investigations to sub-stances possessing the three-carbon t'automerio complex but con-taining groups other than carboxyl attached to its terminal carbon atoms-preferably to some substance in which the possible a 144 INGOZD AND TEOEPE ELIMINATION OF THE sym9let.q of the molecule could be tested without in any way tampering with the three-caqbon system. Such a substance presents itself in the hydrocarbon indene the analogy of which to gluhconio acid becomes apparent when the formulae are written together thus: CO,H*CH 'Z>CH ( '\CTT I '2>CH Uo$?H*'H CO,H*CH C'O,H*CH /'&I I >CH I !,,>,,* \PH I \ I (1.1 (111).(11.1 (IV. 1 It was hoped originally that the accuracy either of I11 or of IV might be proved by preparing solid substituted indenes from a-hydrindones of the types V and VI, If the unsubstituted three-carbon system of the indene nucleus is symmetrical as would be expected from analogy t o unsubstituted glutaconic acid the ketones V and VI should yield the same indene whilst if unsymmetrical two different indenes should result . We have not yet been able to elaborate methods leading to the preparation of suitable ketones of the types V and VI in sufficiently large quantities to ensure the success of this method of attack.In the meantime however we have made use of more easily available materials to obt4ain evidence bearing on the subject and in par-ticular to investigate a reaction which has been repeatedly observed among esters of the glutaconic series and appears to be peculiar to tautomeric compounds. In 1905 the observation was made (Rogerson and Thorpe T., 1905 87 1702) that ethyl y-cyano-a@y-trimethylglutaconate readily passed into ethyl carbonate and ethyl y-cyano-apy-tri-methylcrotonate under the influence of cold sodium ethoxide : CO,Et*CMe(CN)*CMe:CMe*CO,Et + EtOH + CHMe( CN)*CMe:CMe*CO,Et + CO (OE t)2. Since that time fairly extensive use has been made of this reac-tion in the preparation of a series of alkylated glutaconic esters (Thole and Thorpe T.1911 99 2187). Thus the monoalkylated products derived from Conrad and Guthzeit's yellow sodium com-pound (Annulen 1883 222 259) yielded ethyl carbonate along with tribasia esters: (CO,Et),CR*CH:C(CO,Et) + EtOH + CO2Et.C HR CH :C( CO2E t)a + CO( OEt) CARBETHOXYL GROUP PROM T A U T ~ I C SYSTBMB PART I. 145 The tribasic esters did not decxmpme when treated with excess of the same reqent but on further alkylation gave ay-dialkyl deriv-atives which reacted readily to form ethyl carbonate and dialkylahd glutaconic esters : CO,Et*CR:CH*CR’(CO,Eti) + EtOH + CO,Et.CR CH*CHR/*CO,E f + CO (OE t),. The study of these and similar casw led to a generalisation regarding the determining cause of these reactions. They have always been found to be peculiar in substances of the glutaconic hype to those in which all the terminal hydrogen atoms of the three-carbon system have been substituted.It was therefore inferred that the tendency in such cases to acquire the hydrogen atom necessary to enable the substance to pass into its tautomeric form is such that a carbethoxyl group readily becomes detached from the molecule and replaced by an atom of hydrogen under the influence of a suitable reagent. We shall have occasion more than once to make use of this general rule. The ethyl carbonate reaction is therefore very closely bound up with the tautomerim of the three-carbon system. One would not t.herefore expect derivatives of vinylacetic acid such as those represented by formuls V I I and VIII in which the double bond would be purely static to exhibit this reaction to any marked degree.(1 ) Bz*CH .. CH2 C E ~ ~ H CH Ph-C c/GH4-’C;s 6R( CN) C0,Et 6 R( CN)mCO,E f (46R((3N) *COEti (mT. 1 (VIII.) (1x4 It was therefore decided to prepare an ester of the type IX and investigate its behaviour towards Gold sodium ethoxide. The similarity with V I I and V I I I is clear. I f the double bond in the ester IX really possesses the same stable character w0 should for similar reasons expect it to be unreactive. This conclusion is in agreement with the generalisation above cited; for if the three-carbon system (l) (2) (3) (see formula IX) in the indene ring is non-tautomeric and the double bond quite static between the carbon atoms (2) and (3) then since this same double bond enters also into the three-carbon system (2) (3) (4) the latter must be non-tautomeric as well.Its normal form would clearly be incap-able of existence since the central carbon atom (3) is rendered permanently quaternary by the double bond. I n such an eshr we should not in view of the above-mentioned generalisation, expect ta find any tendency to acquire an atom of hydrogen which 146 INGOLD AND THORPE ELIMINATION OF THE if it were acquired could not possibly be mobile. If on the con-trary we found that an ester of the type IX actually did possess a noteworthy tendency to acquire a hydrogen atom in place of its carbethoxyl group we should have to look on the fact as evidence of the tautomeric or dynamic character of the three-carbon system (l) (2) (3) of the indene ring.This follows by simply reversing the argument. Actually we have succeeded in preparing a number of indenyl 3-cyanoacetic esters of the type IX and have found that the lower members of the series possess a very marked ten-dency to lose their carbethoxyl group as ethyl carbonate when treated with quite a small quantity of sodium ethoxide a t 30°. The ester in which R=Me for example when treated with as little as one-sixth of a molecule of sodium ethoxide reacts a t 30° in the course of a few minutes. The yield of the decarbeth-oxylated nitrile is 60 per cent. the remainder of the material passing into an insoluble substance of high molecular weight. I n all the cases of this reaction investigated there was a greater or less quantity of insoluble by-product formed along with the nitrile X and ethyl carbonate./'CH + Et@R -+ (',cH6>CH -I- CO(OEt),. "-CR(CN)*CO,Et CHR-CN \/:- I I @a: (X.1 With homologous alkyl derivatives (R=Et and R=Pr.) the reaction becomes successively more sluggish and an increased quantity of by-product is formed the yield of nitrile consequently diminishing. This is quite analogous to all that has been observed in regard to the same reaction when applied to the glutaconic esters (T. 1911 99 2192). Two points to which no analogy has as yet been investigated or observed among the glutaconic esters, require however special notice. The ester for which R=allyl was found to be very much more reactive than the corresponding n-propyl derivative. I t s reactivity was quite of a similar order to that of the methylated ester.The other point is that a branched chain in the alkyl group appears to inhibit the reaction practically ahgether. Thus the esters for which R was isopropyl, isobutyl and isoamyl gave no nitrile after remaining for twenty-four hours with one-sixth of a molecular proportion of sodium ethoxide a t 30°. In order to obtain a comparative check on these results we decided to investigate an indenyl-2-cyanoacetic ester of the type XI. This clearly differs from IX only in the fact that in XI th CARBETHOXYL UROUP FROM TAIJTOMERIO SYSTEMS. PART I. 147 cyanoacetic residue is attached t o the central carbon atom of the indene system. Now if the double bond in XI is entirely static as regards possible interchange across the system (l) (2) (3) then the carbon atam (4) will be the terminal carbon atom of one three-carbon system only namely the system (3) (2) (4).It should therefore differ but little in reactivity from the carbon atom (4) in the corresponding ester of type IX. If on the other hand the system (l) (2) (3) of XI possesses a mobile hydrogen atom and a mobile double bond it is clear tlhat a hydrogen atom attached to the carbon atom (4) will have a double possibility of '' wandering "; it might wander either to (1) or to (3). We might, therefore in view of the general rule expect to find an ester of the type XI even more prone than the corresponding ester of the type IX to exchange its carbethoxyl group for an atom of hydrogen. Experiment shows the lat'ter supposition to be amply justified.The ester prepared was that for which R=Me. With one-twentieth molecular proportion of sodium ethoxide there was obtained after three minutes atl 1 5 O a practically quantitative yield of the corresponding nitrile (XI1 R =Me). This connexion between the ease of elimination of the carb-ethoxyl group and the potential mobility of the hydrogen attached to the carbon of the cyanoacetic residue suggests a possible ex-planation of the broad facts both in the indene and glutaconic series in regard to the effect of the size of an alkyl group on the ease of the reaction. It seems likely t o be connected with the fact that when heavier alkyl groups were introduced into the gluhconio molecule they were found to increase the stability of the un-saturated form and consequently to reduce the predominance of the normal and the potential mobility of the tautomeric hydrogen atom (since tautmerism depends 0x1 the possibility of the exist-ence of the normal form).An ester in which the degree of tauto-merism of the three-carbon system has been so reduced by the entrance of a large alky1 group would in view of the generalisa-tion be expected to exhibit a smaller tendency t o acquire an a b m of hydrogen and this is what is actually found to be the case. I n complete accord with the great ease with which the ester XI exchanges its carbethoxyl group for an atom of hydrogen and with the presumed excessive mobility of the latter is the behaviour o€ the unmethylated ester XIV. The esters IX and XI wer 148 IRGOLD AND THORPE ELIMINATION OF THI obtained by alkylating the esters XI11 and XIV r e a p t i d y .These esters differ in acidity in the sense that as one would expect from the different reactivities of t,heir alkyl derivatives the cyano-acetic hydrogen shm of indenyl-2-cyanmetic eater (XIV) is more loosely attached than that of the corresponding indenyl-3-deriv-ative (XIII). Thus ethyl indenyl-2-cyanoacetate is a weak acid I A C H I 2>C*CH(CN)*C0,Et @Y(@CH 6H( CN)*CO,Et \PH (XIII.) (XIV.) forming a sodium salt which is not hydrolysed in aqueous solution, being decomposed only by slightly acid substances such as carbonic acid. Ethyl indenyl-3-cyanoacetate on the other hand only forms a sodium salt in complete absence of water. Jt is perhaps worOh noticing that when either of the indenyl-cyanoacetic esters (XIII and XIV) are converted into or liberated f ram their salts a deep crimson colour is immediately developed.This fades in the course of a few seconds both the free esters and the solid salts being colourless. Another colour change which was regularly observed in the course of these experiments took place when the alkylated esters (IX and XI) were treated with sodium ethoxide. An indigo-blue colour immediately developed and gradually faded as the elimina-tion of the carbethoxyl group proceeded. The preparation of the indenylcyanoacetio esters (XIII and XLV) was readily accomplished by condensing a-hydrindone or 8-hydrindone with ethyl cyanoacetate in the presence af piperidine or diethylamine : +CH,(CN)*CO,Et + \ ’kB”>CA +H,O bH(CN)*CO,Et \/--c When ethyl indenyl-3-cyanoacetate was hydrolysed either by acid or by alkali the cyano-acid (XV) was formed although not without considerable decomposition.This acid on heating above its melting point gave off carbon dioxide and from the dark UARBETHOXYL GROUP FROM TAUTOMERIO SYSTEMS. P U T I. colourad residue indenyl-3-acetonitrile (XVI) wa.s isolahd by 1 ''Cb>~R !-c \/ ()"Hp>m CH,-CN \/-c ~IH(CN).CO,H (XVJ (XVZ.) vacuum distillation. This is the first member of the. series of homologous nitriles of which X is the type. It cannot of course, be prepared directly from the cyano-ester (XIII) by the action of sodium ethoxide for reasons already indicated. Neither the compound XVI nor any of its homologues appears to form a sodim compound when treated with alcoholic sodium ethoxide and all attemph to introduce another alkyl group into these compounds using sodium or potassium ethoxide and an alkyl iodide met with failure.The same was the a s e when the methylatd nitrile XII derived from j3-hydrindone was used. E X P E R I M E N T A L . The a-hydrindone required for these experiments was prepared f m j3-phenylpropionic acid by a met4hod essentially the same as that described by Kipping (T. 1894 65 680) but with the intro-duction of certain modifications which so improved the yield as ta make this substance far more easily available than it has hitherto been. &PhentyZpr@onyE Chloride .-It was found advantageous to use thionyl chloride in the preparation of this substance instead of phosphorus pentachloride.8-Phenylpropionic acid (100 grams) was mixed with an equal weight of thionyl chloride in a flask fitted with an efficient condenser. The reaction was started by gentle heat and allowed to proceed for one and a-half to two hours when the evolution of gas had ceased. The contents of the flask were then transferred t o a Claisen distillation flask and heated a t looo/ 25 mm. until all the thionyl chloride had distilled over. The resi-due was then fractionated under 22.5 mm. pressare and 110 grams boiling a t 121-122O were collected. The theoretical yield is 112 grams. a-Hy&ndone .-Pure 8-phenylpropionyl chloride being thus available it was found possible to carry out the internal con-densation whereby hydrogen chloride is eliminated and cthydr-indone produced with much better results than Ripping was able to obtain with the impure chloride at his disposal Whilst he seldom obtained more than a 56 per yield it was found that with the pure chloride a yield of 75 per cent..was always secured. Th 150 INUOLD AND THOBPE ELIMINATION OF THE reaction with the pure chloride is far more violent than wihh the impure product and hence the mixture must be heated for a few minutes only With this exception the details given by Hipping were closely followed. &Hydrindme.-The 8-hydrindone required for these experi-ments was prepared by the improved modification (P. 1911 27, 108) of the original process described by Moore and Thorpe (T., 1908 93 165).Condensatiom of a-Hydrindone with Ethyl Cynnoacetate in the Presence of Secondary Rases Ethyl Indenyl-3-cyanoacetate (XIII p. 148). Since a-hydrindone readily dissolves in ethyl cyanoacetate i t is not necessary t o use any solvent in this condensation. A solution of 19 grams of the ketone in 16 grams of the ester was treated with 6.5 grams of diethylamine and the mixture allowed to remain a t 40° for twenty-four hours. A t the end of that t h e the tube, which contained a stiff paste of crystals of the condensation pro-duct was cooled for an hour a t Oo and the crystals were drained on porous porcelain. The compound separates from alcohol in colourless needle-shaped crystals melting a t 104O ; it is moderately soluble in dry ?ther and readily so in benzene chloroform or acetone.The yield represents about 55 per cent. of the theoretical, and is but little affected when piperidine is used in place of diethylamine : 0.1031 gave 0.2805 CO and 0.0538 H,O. 0.2492 , 13.8 C.C. N2 a t 19O and 742.6 mm. N=6*19. C,,H1,O,N requires C = 74.0 ; H = 5.7 ; N = 6.2 per cent. The ester reacts with alcoholic sodium ethoxide forming a sodium compound from which the ester is regenerated by the action of water. There is no doubt but that this sodium com-pound contains the metal attached to the cyanoacetic residue and that therefore the ester described above has the constitution assigned to it,. When alcoholic sodium ethoxide was added to the ester a deep crimson colour was invariably formed. This faded after a few seconds to a bright yellow which persisted so long as the solution remained alkaline.C=74.20; H=5.80. Condensation of a-Hyd&done with Ethyl Cyanoacetate in the Presence of A ZcohoJic Sodium Ethoxide. The condensation with sodium ethoxide appears to be of a con-siderably more complex character than when secondary bases are used. Thus when an alcoholic solution of a-hydrindone is adde UARBETHOXYL GROUP FROM TAUTOMERIC SYS!t'EMS. P U T I. 151 to a hot suspension in alcohol of the s o d i m compound of ethyl cyanoacetate there is formed a mixture of substances which may be precipitated by adding water. This mixtnre consists chiefly of two compounds melting a t 143O and 88-89O respectively which may be separated and obtained in a s t a h of purity by fractional crystallisation first from alcohol and finally from a mixture of absolute alcohol and benzene.The former compound was identi-fied with anhydrobis-a-hydrindone (Found C = 87-74 ; H = 5.70. Calc. G=87*8; H=5-7 per cent.) which is recorded as melting a t 142-143* (Kipping T. 1894 65 495). Ethyl 2 3f-Di-indenyl-3-cyanoacetate, /-\ '\ / I /"CH 1 2>C*C/-j \/-c ~ H - C H , bH(CN)*CO Et The substance melt5ng a t 88-89O may be made to become the principal product if the order in which the condensing substances are mixed is reversed 3.3 Grams of a-hydrindone were dissolved in a small quantity of hot. alcohol and a hot solution of 0.6 gram of sodium and 2.8 grams of ethyl cyanoacetate in 15 grams of alcohol was slowly added. A few minutes after the addition was complete the solution was rapidly cooled and poured into water.Hydrochloric acid was then added and the oily precipitate ex-tracted with ether the extract washed with dilute sodium carbonate solution and with water and then dried. The solid residue obtained on evaporation of the ether when recrystallised from alcohol weighed 0.8 gram: C=80*78; H=5*68. 0.1545 gave 0.4575 CO and 0-0790 H,O. 0.1818 , 6.7 C.C. N2 a t 17O and 766.1 mm. N=4*23. Ethyl 2 3fdiindenyl-3-cyanoacetate separates from the usual solvents in pinkish-buff needles melting a t 88-89O. It is oxidised instantly by cold alkaline pemanganate. With alcoholic sodium cthoxide it forms a yellow sodium compound from which the original ester can be regenerated. C,H,,O,N requires C= 80.9 ; H = 5.6 ; N = 4.1 per cent.liydrolysis of Ethyl ZnuEenyl-3-cyanmcetate Znc2enyl-3-cyanoacetic Acid (XV p. 149). The hydrolysis of the ester melting a t 104O is a matter of some difficulty owing to the ease with which it undergoes deep-seate 152 1;BQOLD AND TAOECPE ELIMINATION OF THE decomposition with acids and alkalis. Thas on boiling with acids (dilute hydrochloric or sulpburic) only a 6 per cent. yield of the acid is obtained. The acid can be produced in 36 per cent. yield by alkaline hydrolysis but only by working within very narrow limits. Four grams of the eshr were treated with 8 C.C. of 4N-sodium hydroxide and the mixture was heated as rapidly as possible to the boiling point and maintained there for twenty seconds with vigorous shaking. The oil dissolved forming a clear red solution which was kept boiling for thirty seconds longer and then rapidly cooled.The crystalline sodium salt which separated was collected dissolved in water and the solution after passing hhrough a wet filter acidified with hydrochloric acid. The acid separated as a white precipitate which crystlallised from alcohoI in small prisms melting and decomposing a t about ZOOo the melt-ing point depending on the rate of heating. The point of instant-aneous decamposition as measured by the Maquenne block is 2 3 7 O . The acid is sparingly soluble in water or dry ether : 0*1251 gave 0.3326 CO and 0.0509 H,O. 0.2164 , 13.8 C.C. N a t 19O and 755 mm. N=7.24. C=72*51; H=4-52. Cl,H90,N requires C = 72.4 ; H = 4.5 ; N = 7.1 per cent. Indeny I- 3 -ace t onit r il e (XV I p .1 49). The pure recrystallised acid (4.4 grams) was heated a t 250° until the evolution of carbon dioxide had ceased. The dark-caloured oil which remained was then distilled under diminished pressure and the colourless distillate cooled in ice. The solid residue which melted below the ordinary temperature was recrystallised from light petro,leum below Oo and obtained in long colourless needles melting ate 18O : 0.0820 gave 0.2568 CO and 0.0426 H,O. 0.1430 , 11.3 C.C. N a t 19 and 761.2 mm. N=9.06. CI1H,N require6 C=852; H=5*8; N=9.0 per cent. The attempts which were made t'a alkylate this nitrile did not meet with any success and we were quite unable to find the con-ditions by which the nitrile could be hydrolysed to the correspond-ing acid.C=85*41; H =5*77. ,4 lkyhtion. of Ethyl IndenyyL3-cyanoacetate and the Elimination of the Cwbethozvl Group Ethd a-lndemyl-3-a-cyanlo-I n order to prepare this subskance 12 grams of the ester melt CARBETHOXYL QROW FROM TAUTOMERIC! SYSTEHS. PART I. 153 ing a t 104O were dissolved in the least possible quantity of alcohol a t 70° and added to a solutiun of 1.2 grams of sodium in 16 grams of alcohol. Ten grams of methyl iodide were then added and the mixture was heated until the yellow colour had entirely dis-appeared and the solution had b e m e neutral an operation which usually required ten minutes. The addition of w&r precipibted an oil which when extracted by ether yielded a solid residue after the solvent had been evaporated.The compound crystallises from a mixt'ure of light petroleum and ether in large cubes melting at 60°; it is readily soluble in the usual organic solvents excepting light petroleum. The yield was 70 per cent. of the theoretical: 0.1418 gave 01.3900 CO and 0.0816 H,O. 0.2818 , 14.4 C.C. N a t 19O and 783 mm. N=5-92. C=74.96; H=6.39. @15H&N requires C = 74.7 ; H = 6.2 ; N= 5.8 per cent. Six grams of the carboxylic ester were dissolved in cold alcohol and an alcoholic solution containing 0.1 gram of sodium was added. The solution was kept at' 30° for a short time when the blue colour which had developed was discharged and the liquid had a strong odour of ethyl carbonate. The liquid was poured through a filter, water was added and the precipitate which was formed was induced to solidify by shaking.It' was then collected dried and extracted with hot light petroleurn the nitrile being deposited from the solvent on cooling in long colourless needles melting a t 118O. It may also be recrystallised from dilute alcohol. The yield is 60 per cent. of the theoretical : 0-1032 gave 0.3227 CO and 0.0603 H,O. C=85*28; H=6*49. 0.2118 , 15-4 C.C. N a t 2 3 O and 771 mm. N=8-30. The nitrile could not be hydrolysed and all attempts to i n t r e (&HI1N requires C =85*2 ; H= 6.5 ; N = 8.3 per cent. dnce another alkyl group into it were without success. Ethyl a-ZndengL3-a- cymo-n- b utyrat e , >C* C Et (CN )* C0,Et. PH,*CH C,H,-This ester was prepared in the same way as the methyl derivative It is a colourless oil which boils a t ZOOo/ already described.25 mm.: 0.1304 gave 0.3604 CO and 0.0785 H,O. 0.2363 , 11.6 C.C. N a t 2 2 O and 768 mm. N=5*62. C=75*38; H=6.69. C,,R,,O,N requires C=75-3; H=6*7; N = 5 3 per cent 154 MGOLD AND THORPE ELIIKISATZOX OF THE a-Zndemy l-3-n- b ut yronitril e ?H2'cH>C*CH Et*CN. c,*,-This nitrile was produced from the carboxylic ester by the action of a small quantity of alcoholic sodium et,hoxide under the same conditions as those which were described for the methyl derivative. The crude solid precipitated by water was extracted with hot alcohol and the nitrile obtained from the alcoholic extract by the addition of water. It crystallises from light petroleum in long needles melting a t 76O. The yield is 20 per cent. of the theoretical : 0.1306 gave 0.4067 CO and 0.0835 H,O.0.1882 ,) 12.6 C.C. N2 at 23O and 771 m. N=7*64. CI3Hl3N requires C=85*2; H=7*1; N=7.7 per cent. C=84*93; H=7.11. Ethyl a-lndenyl-3-accyan~o-n-ua~e~atc, ~H2BCH>C*CPru(CN)*C0,Et. C,H,-This ester was produced by the acttion of n-prop91 iodide on the sodium compound of ethyl indenyl-3-cyanoacetate in alcoholic solution. The reaction was complete after heating for forty-five minutes and the product was then isolated in the usual way. The ester is an oil which boils a t 21Oa/2O mm. : 0.1259 gave 0.3502 CO and 0.0797 H,O. 0.2169 , 10.0 C.C. N a t 2 2 O and 768 mm. N=5-26. C=75*86; H=7*03. C,,H,,O,N requires C = 75.8 ; H = 7.1 ; N = 5.2 per cent. This compound was prepared in the same manner as the ethyl derivative although in the present instance the reaction proceeded much less readily.It was isolated in the usual way and crystal-lised from light petroleum in cdourless needles melting a t 67O. The yield was only 10 per cent. of the theoretical: 0.1028 gave 0.3210 CO and 0*0710 H,O. 0.2018 ,? 12.8 C.C. N a t 23O and 768 mm. N=7.20. C=85.16; H=7-68. C,,H,,N requires C=85*3; H=7*6; N=7*1 per cent. Ethyl a-Zndenyl-3-a-cyanoisovalera t e , > C CPr@( CN) CO Et . vH2*CH C6E4--isoPrapyl iodide was found to react with the sodium compoun CARBETHOXYL GROUP PROM TAUTOMERIC SYSTEMS. PART I. 155 of ethyl indenyl-3-cyanoacetate in the same manner as n-propyl iodide and the product was isolated in the 8ame way. In this case the ester was obtained as a colourless oil which boiled a t 260°/ 120 mm.and solidified in the receiver. The solid crystallised from light petroleum in colourless prisms melting a t 72O. The yield represented 60 per cent. of the theoretical : 0.1115 gave 0.3105 CO and 0,0716 H,O. 0.1859 ) 8.6 C.C. N a t 22O and 768 mm. N=5*28. Cl7Hl9O2N requires C = 75.8; H = 7.1 ; N = 5.2 per cent. This ester was scarcely changed by alcoholic sodium ethoxide under the experimental conditions which caused the other esters to lose their carbethoxyl groups as ethyl carbonate. Most of the original ester and a small amount of insoluble matter were C=75-95; H=7*15. recovered. Ethyl a-Zndenyl-3-a-cyanoallylacetat e, >C*C(CH,*CH:CH,)( CN)*CO,Et. C H,*CH b6H4-Ally1 iodide reacted with the sodium compound of ethyl indenyl-3-cyanoacetate in boiling alcoholic solution in the course of a few seconds.The product was isolated in the usual way and crystal-lised from light petroleum containing a little dry ether in nearly cubical crystals melting a t 65O. The yield was 65 per cent. of the theoretical : C=76-63; H=6*41. 0.1240 gave 0~3484 CO and 0.0715 H,O. 0.2954 )) 13.6 C.C. N a t 19O and 764 mm. N=5*31. C17H1702N requires C = 76.4; H = 6.4 ; N = 5.2 per cent, a-Zndeny Z-3-ally lace t mit rile ) yIJ2'CH>C*CH(CN) *CH ,* CH,:CH,. C@4-The action of a trace of alcoholic sodium ethoxide on the ester caused the carbet'hoxyl group to be eliminated and gave a yield of 40 per cent of the corresponding nitrile the same conditions being employed as those described in the former experiments.The nitrile crystallises from light petroleum in colourless needlm melt-ing a t 108O: @=86.28; H=6.75. 0.1064 gave 0.3366 CO and 0.0647 H,O. 0.2732 ,) 17.6 C.C. N2 a t 23O and 768 am. N=7*31. C,,H,@ requires C=86*2; H=6*7; N=7.1 per cent 156 INGOLD AND TBORPE ELIMINATION OP THE This ester was prepared in the usual manner from imbutyl iodide It distilled a t 260°/40 mm. as C=76.43; H=7-33. and ethyl indenyl-3-cyaaoacetate. a pale yellow oil : 0.1956 gave 0.5481 CO and 0.1290 H,O. 0-2600 , 11-6 C.C. N a t 22O and 761 mm. N=5.07. Like the isopropyl derivative this ester did not lose its carb-ethoxyl group by treatment with cold sodium ethoxide. After being submitted t o the same experimental conditions as the other esters the recovered material gave on analysis C = 76.89 H = 7.56, N = 5.13 indicating that it was practically unchanged.(The de-carbethoxylated compound Cl,H17N requires C = 85.3 ; H = 8.1 ; N=6.6 per cent.) Cl,H,102N requires C = 76.3 ; H = 7.4 ; N = 4.9 per cent. Ethyl a - l d e n yl-3-a-cyanoiso hept oat e, When prepared from the ester melting at 104O and isoamyl iodide and isolated in the usual way this ester distilled a t 270"/ 34 nun. as an almost colourlees oil: 0.1680 gave 0.4749 CO and 0.1173 H,O. 0.2127 , 9.0 C.C. N at 22O and 766 mm. N=4.80. The carbethoxyl group could not be eliminated under the customary experimental conditions. The material recovered from the solution of sodium ethoxide gave on analysis C=77.31, H=7*92 N=4*95 indicating that it consisted of the unchanged compound (ClsHlsN the carbethoxyl-free compound requires C=85*3 H=8.5 N = 6 - 2 per cent.).C=77-08; H=7-76. C19H2302N requires C = 76.8 ; H = 7.7 ; N =4.7 per cent. Condensatiom of #3-€€y&h&me with Ethyl Cyanoacetate in the Presence of Secondary Bases. When a mixtiure of B-hydrindone and ethyl cyanaacetah is treated with a secondary base such as piperidine or diethylamine, there is generally formed a mixture of two crystalline compounds melting at 116O and 176O respect.ively. The latter contained no nitzogen and gave on analysis C = 87.62 H = 5.81 (CIBHl,O requires C=87-8; H=5.7 per cent.). It is therefore probabl CARBETHOXYL QROUP FROM TAUTOMERIO SYSTEMS. PART I. 157 identical with anhydrobis-P-hydrindone the melting point of which is given as approximately 170° (Heusler and Schieffer Ber.1899, 32 32). The amount of bis-compound formed varies very much with the conditions and unless the condensation is kept well under control it may become the sole product. Ethyl Zr.demyl-2-cyanoacetate (XIV p. 148). By exercising care it was found possible to obtain a solid pro-duct containing as much as 65 per cent. of ethyl indenyl-2-cyano-acetate and 35 per cent. of anhydrobis-B-hydrindone. Ten grams of P-hydrindone were dissolved in 9 grams of ethyl cyanoacetate, and the solution was cooled below 1 8 O while 30 drops of diethyl-amine were added. After the addition of each drop the solution was immediately shaken and well cooled in running water. After completing the addition of the base the tube containing the mix-ture was immersed in cold water for thirty minutes when it was withdrawn and allowed t o remain a t the ordinary temperature for forty-eight hours.At the end of that time the stiff paste of crystals which filled the tube was spread on porous porcelain and allowed to remain until colourless. The crude solid mixture of condensation product.s which usually weighed about 13 grams was rubbed to a fine powder under a little dry ether and roughly separated by extracting with four times its weight of boiling 95 per cent. alcohol the bulk of tche bis-compmnd being left un-dissolved. The crude ester deposited by the filtrate melted between 90° and l l O o . It was finely powdered and stirred into an excess of 4N-sodium hydroxide a t 30° the whole diluted with an equal bulk of water quickly filtered and t'reated with aqueous sodium hydrogen carbonate in excess.The precipitated ester was caused t o solidify by shaking and then collected and triturated with water. After draining and rwrystallising from alcohol it was obtained in long colourless needles melting a t 116O: 0.1261 gave 0.3417 CO and 0.0646 H,O. 0'1834 , 9.7 C.C. N a t 169 and 772-5 mm. N=6.25. The compound is very readily soluble in hot alcohol but apar-ingly so in cold. It is also very readily soluble in cold benzene, chloroform or acetone and sparingly so in ether or light petroleum. It tends to form coloured products when its alkaline solution is exposed t o the air and the yield obtained by the sodium hydroxide separation t'heref ore depends greatly on the speed with which the operations are Carrie out.The separation was also effected by means of a long series of frac-VOL. CXVI. C=73*91; H=5-70. C,,H,,O,N requires C = 74.0 ; H = 5.7 ; N = 6.2 per cent 158 ELIMINATION OF THE CARBETHOXYL GROUP ETC. PART I. tional crystallisations from alcohol. The ester obtained by both methods proved to be the same .substance showing that the form-ation of a sodium salt had not involved any isomeric change and that the compound must therefore have the structure assigned to it. The ester is readily soluble in 4iT-sodium hydroxide and is not reprecipitated when a large bulk of water is 'added. It is insoluble, however in sodium carbonate and is therefore precipitated from the hydroxide solution by carbon dioxide or a bicarbonate.During the precipitation by either of these reagents or by an acid a transient red colour always appeared. A similar transient colour was invariably observed when an alcoholic solution of the ester was treated with aqueous or alcoholic potassium hydroxide or alcoholic sodium ethoxide. Sodium Bem'vative.-One gram of the ester was dissolved in twice the theoretical quantity of 4J-sodium hydroxide a t 50°. On cooling a colourless crystalline sodium derivative separated out. The alkaline liquid was poured off from the crystals which were then washed with ice-water and dried in a vacuum over phosphoric oxide : 0.3002 gave 0.0860 Na,SO,. C,,H,O,NNa requires Na = 9-24 per cent. When kept in a closed space the sodium compound slowly decom-poses acquiring a green colour but if spread in a thin layer over a large area in a dry atmosphere it can be kept for several weeks.Although the compound itself is colourless its solution in water is orange. This solutioa on acidification becomes deep red for a few moments the colour quickly fading as the free est'er separates out. Na = 9-28. Met h y lntion of E't h y 1 Indenyl- 2-c yanoac e t at e and the Elimination of the Carbethoxyl Group Ethyl a-lndenyl-2-a-cyamo-propio na t e C H,<(g> C CM e ( CN ) C 0 I! t . The methylation of ethyl indenyl-2-cyanoacetate was accom-plished both by the action of methyl iodide on the dry sodium campound suspended in alcohol and by the more usual process of treating the free ester with alcoholic sodium ethoxide and methyl iodide.The est'er was precipitated with water and extracted with ether. After washing the extractl with water and drying the ether was evaporated and the residual oil crystallised from light petroleum containing a t'race of ether. The ester separated in dense colourless prisms melting a t 5 6 O . The yield was about 70 per cent.. of the theoretical PREPARATJON OF MONOMETHYLAMINE FROM CHLOROPICRIN. 169 0.1035 gave 0.2837 CO and 0.0580 H,O. 0.1653 , 8.4 C.C. N a t 1 8 O and 779 mm. N=5.90. C=74-23; H=6.22. C,,H,,O,N requires C= 74.7 ; H = 6.2 ; N = 5-8 per cent. The elimination of the carbetholxyl group of ethyl indenyl-2-cyanopropionate was found to proceed with great ease in the presence of a small quantity of sodium ethoxide. Thus with one-twentieth of a molecular proportion of sodium at; 1 5 O the reaction was complete in about three minutes. On adding water the nitrile separated out. After allowing the suspension to remain for twenty-four hours it was filtered and the solid dried and recrystal-lised from light petroleum from which it separated in long colour-less needles melting a t 92O: 0.1011 gave 0.3167 CO and 0.0591 H,O. 0.1179 , 8.5 C.C. N a t 20° and 764 mm. N=8*22. The yield was practically quantitative. All attempts t o hydrolyse this nitrile resulted in deep-seated decompositions taking place and we were unable to isolate the corresponding acid. Several attempts also were made to introduce anot?her methyl group into the molecule but without success. C=85*43; H=6-49. CI2H,,N requires C=85*2; H=6*5; N=8.3 per cent. IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY, SOUTH RENSINGTON [Received December 6th 191 8.