282 CURTIS AND KENNER:XXXI. -The Condensation of Ethyl Glutaeonate.By RAYMOND CURTIS and JAMES EENNER.THE formation of act- and of ay-dirnethyl glutaconates by the directmethylation of ethyl glutaconate (Henrich, Mormtsh., 1899, 20,539; Blaise, Compt. r e d . , 1903, 136, 639) suggested to us thepossibility that by the interaction of ethylene dibromide and ethylglutaconate in the presence of sodium ethoxide, either a cyclo-propane (I) or a cyclopentene (111) derivative, or both of thesemight be obtained. After our experiments in this direction hadbeen completed, we discovered that Fecht (Ber., 1907, 40, 3883THE COSDEKSATION OF ETHYL GLUTACONATE. 283had previously studied this reaction, and embodied his results in apaper, the title of which would not suggest that such experimentsmight be described in it.This author obtained evidence of theformation of the cyclopropane derivative (I) by isolating the esterof its reduction product (IT) :7H2 C0,Et + CH2>C<CH2*CH2*C02Et;(11.1CH,.QH*CO,EtCHq'C*CO,Et(111.)I EHNo other compound is mentioned, although i t is evident fromthe account given, as well as from a consideration of the experi-ments now described, and of those of Blaise (Zoc. c i t . ) on thealkylation of ethyl glutaconate, that only a very small amount ofcyclopropane derivative could have been formed; the fate of themajor portion of the ethyl glutaconate theref ore remained unknown.The experiments carried out in the present instance furnished awhite, crystalline solid, and some oily matter (10 per cent.of theyield). The latter, although formed in such small amount, wasprobably similar in its nature to the product obtained by Fecht,and has not yet been further investigated. The crystalline com-pound gave a blue colour with ferric chloride, and had acidicproperties (equivalent = 316). This, together with its composition,pointed to its production from two molecules of ethyl glutaconate(mol. wt. = 186) by the elimination of the elements of one moleculeof alcohol, the ethylene dibromide having taken no part in thereaction.It was then found that the properties of two condensationproducts of ethyl glutaconate, which had already been described,were in agreement with those of the compound isolated by us.On the one hand, Blaise (loc.cit., Bull. SOC. chim., 1903, [iv],29, lola), by heating an alcoholic solution of ethyl glutaconatewith sodium ethoxide a t looo, isolated it compound to which heassigned the constitution of ethyl glutaconylglutaconate (IV) :QH2*CH :CH*CO,Et ---+ SH,*CO,Et + CH:C H*C02Et C0,Et$?H,--CO---$!H*CH:CH*CO,EtCH :CH*CO,Et C0,Et(IV.)On the other hand, Pechmann, Bauer, and Obermiller (Ber.,1904, 37, 2113) showed that ethyl glutaconate undergoes con-u 284 CURTIS AND KENNER:densation when its ethereal solution is heated with sodium ethoxide.From analogy to Pechmann'~ work on the formation of ethyla-methyleneglutarate from ethyl acrylate under similar conditions,the reaction in the case of ethyl glutaconate was expressed in thefollowing manner :70,Et 70,EtQHz 70,Et 7H2 yO,EtGH + GH + FH-E +YH p a yHqo,I!:t ?O,Etp 2 7%CO-CH*CO, Et co-bH,C0,Et CH,*CO,Et c.'" CH,.C0,Et C0,Et$!H--E.CO,Et or 7H-q. C0,Etp 2 p EtO,C*QH CHVn.) (Vb. )This view as to the course of the reaction received confirmationfrom the oxidation of the compound with bromine and the ultimateconversion of the product into l-hydroxy-2 : 4-dicarboxyphenyl-acetic acid, but no decision could be reached as to which of thealternative formulae (Va and Vb) represented the original condensa-tion product.In spite of some discrepancies between the accounts given of thesetwo products, it appeared to us highly probable that they wereidentical with each other and with our product. A repetition ofthe experiments described by Blaise and by Pechmann, respectively,was therefore undertaken in order to decide this point, and toexamine the outstanding differences.It was found that all three substances were identical, for they(1) melted a t very approximately the same temperature,* and noneof the compounds depressed the melting points of the others;(2) were converted into the same benzene derivative by oxidationwith bromine; and (3) furnished the same condensation productwith phenylhydrazine.Further, the condensation takes place inaccordance with the view of Pechmann and his co-workers, and the* The slight differences in melting point are due t o traces of other products whichare difficult to remove. Thus, if a hydrogeii atom should migrate from eachmolecule of ethyl glutaconate to the other, in the manner represented byPechmann, a cyclobutane derivative would result, aud this appears to be the caseunder certain conditions (compare Pechmann, Ber., 1899, 32, 2301 ; Gutzeit, Ber.,1901, 34, 678).Further, Blaise showed that during the methylation of ethylglutaconate, a portion of the ethyl " glutaconylglutaconate " also underwentmethylationTHE CON DENSATlON OF ETHYL GLUTACONATE. 285compound in question was the chief product in the attemptedcondensation of ethyl glutaconate with ethylene dibromide.Of the two possible formule (Va and Vb), the latter, whichrepresents (in its enolic form) ethyl 2 : 6-dicarbethoxy-A2:~-cyclo-hexadien-5-oZ-luceta~te, appears to us the more probable, for thefollowing reason.The intermediate compound, being a derivativeof ethyl glutaconate, will form a sodium salt, from which theformation of a condensation product would be expected to takeplace in the following manner :yH,-CO,'Et 7 H,*CO,E tf: H-E*CO,E t $! H *$*CO,EtEtO,C*CH, YH --3 EtO,C*QH QHCO-CH, NaO (E t 0) C : C HAccepting this inference, the nomenclature and formule of thevarious derivatives prepared from the condensation product byPechmann and his workers (Zoc. cit.) require to be amended accord-ingly.Pechmann and his collaborztors were unable to prepare a con-densation product of the compound with phenylhydrazine, whilstBlaise was undecided as to whether the derivative isolated by himwas a hydrazone (VI) or a pyrazolone (VII).It has now beenfound that the latter formula is disproved by the inability of thecompound to enter into salt-formation, and that it is to be lookedon rather tics a hydrazide (VIII) than as a hydrazone, for it iseasily oxidised in alcoholic solution by mercuric oxide to a redazo-compound.?El2* CO,E 5CHyH,*CO,EtCH/\\/f! 0-YH g*CO,EtC'H,/\\/(TI.)EtO,C*yH G-CO,EtC,H,*NH*N:C CH C,H,*N-N:C CH(VII.)OH,$!H,*CO,EtCHEt02C*lc;( G*CO,Et/\vCH2C6FiZ,*NH*NH*C CR(VIII.)We found, however, that a pyrazokone derivative was a t onceproduced by the action of hydrazine hydrate on the condensatio286 CURTIS AND KENNER:product ; the carbethoxy-groups were also attacked, so that theresulting compound had the structure IX or X :$lH,*CO*NH*NH, YH,*CO*NH*NH,CH CH/\ /\ $lO--FH g*CO*NIF*NH, yO---E i;‘*CO*NH*NH,NH-N:C CH NH*NH*C CHBy hydrolysis with dilute mineral acid, as described by Blaise,the original condensation product furnished a monobasic acid,C8HI2O4, from which a lactone could be obtained.Although Blaisewas unable to ascribe a constitutional formula to the product, itis evident that its oxidisability by bromine and its easy lactonisationfind expression in the formula of a cyclohexanone derivative (XI) :~E€,-CO,E;t 7 H ,*CO,H FH2*C02HCH CH/\\/EtO,C*Yt-I R*C02Et --+ HO,C*$!H 5!(0€1)*CO,H -+- YH, YH*OHCO CH, I GH2 ; CH,/\\/CH,CO CHc H2(XI. 1From a coi~sideratiou of Pechmann’s explanation of the con-densation of ethyl glutaconate, it seemed probable that a similarcondensation could occur to ethyl a-ethyl glutaconate, but not toethyl ay-diethylglutaconate, whilst the view advocated by Blaisewould permit of both condensations. It was found, however, that,unlike ethyl glutaconate, its a-ethyl derivative did not undergocondensation under either of the conditions employed by Pechmannand by Blaise.An explanation of the difference between ethylglutaconate and its a-substituted derivative may be based on theviews expressed by Bischoff as a result of his researches on chainformation, for i t will be evident that if two molecules of ethyla-ethylglutaconate condense in the manner above suggested, therewill be a much larger number of groups in Bischoff’s critical 1 : 5-and 1 : 6-positions than in the case of ethyl glutaconate itself.Weprefer, however, to place another interpretation on this result;thus, ethyl glutaconate takes part in the initial polymerisation inthe form of its sodium salt, the constitution of which is expressedby the formula XI1 :CO,Et*CH:CH=CH:C(ONa)*OEt( S I I . )CO,Et*OEt:CH*CH:C(ONa)*OEt.(XI 11.THE CONDENSATIOK OF ETHYL GLUTACOKATE. 287I n this formula the hydrogen atom which migrates during theinitial polymeri6ation is shown in heavy type, and it appears prob-able that polymerisation would not take place if this atom werereplaced by an ethyl group, its shown in XIII. If, however, wetherefore adopt this formula for the sodium salt of the ethyla-ethylglutaconate in question, i t follows that the salt-formationtakes place in the carbethoxy-group of this ester remote from theethyl group, a deduction which is in harmony with the conclusionarrived a t by Thorpe and Wood (T., 1913, 103, 1754) on othergrounds.Nevertheless, it is worthy of note that the hydrogen atom, themobility of which is responsible for the polymerisation, is differentfrom the mobile hydrogen atom of ethyl glutaconate, with whichthe recent work of Thorpe and his collaborators has been con-cerned.Pechmann’s reaction is, however, exactly analogous to thereaction by which ethyl sodiocyanoacetate condenses with ab-un-saturated esters, as is evident from the equation representing thisreaction* (Thorpe, T., 1900, 77, 932):CO,Et*GH:CHR + CO,Et*CH:C(ONa)*OEt=CO,Et*CH,*CHR*C( C0,Et) :C( ONa) *OEt .It thus appears that a hydrogen atom attached to an ethyleniccarbon atom does possess a certain mobility if a negative group bealso attached t o this atom.This mobility is perhaps less marked inthe case of ethyl glutaconate than in the case of ethyl cyanoacetate,for whereas ethyl sodiomethylcyanoacetate takes part in the reactionjust mentioned, it has already been explained that ethyl aethyl-glutaconate does not polymerise.* It must be observed that this condensation differs essentially from the ordinaryMichael reaction, in which ethyl sodiomalonate is employed, for in the latter case itis the sodium atom, and not the hydrogen ntoin, that migrates. Thus ethyl sodio-methylmalonate nnd ethyl a-methylacrylate ultimately furnish ay-dimethglglutaricacid (hnwers arid Kobner, Ber., 1891, 24, 1927), whereas ethyl sodiomethylcyano-acetate and ethyl &l-dimethylacrylate yield ethyl y-cyano-413-trimethylbutyrate(Thorpe, Zoc. cit.).The contrast between the sodium salts.of ethyl malonate andethyl cyanoacetate thus illustrated is not confined to this reaction (compare Thorpe,Zoc. cit.), and suggests that they differ in constitution. It seems probably that ethyl,OEtsodiomalonate is more adequately represented by the formula CH(C0,Et):C\ON?....................................(compare Hantzsch’s formula for ethyl sodioacetoacetate, Bey., 1910, 43, 3053)288 CUKTlS AND KENNER:EXPERIMENTAL.Interaction of Ethyl Glutaconate, Ethylene Bihromi.de, and SodiumE t h oxid e.Ethylene dibromide (6.3 grams) was added to a mixture of ethylglutaconate* (6.2 grams) with a cold solution of sodium (0.8 gram)in alcohol (12 c.c.).The mixture remained at the ordinary tem-perature for- several hours, and was finally heated f o r four hourson the water-bath. The product, isolated in the usual manner,consisted of an oil, from which crystals quickly separated. Themixture decomposed when the attempt was made t o distil i t underdiminished pressure,? and the separation of the solid from the oilwas therefore accomplished by filtration. By repeated crystallisationfrom methyl alcohol, white needles were obtained, which melted at78'5O :0.1511 gave 0.3255 CO, and 0.0950 H,O. C=58.76; H=7-05.Equivalent = 317.0*1840 required 5.8 C.C. -W/lO-NaOH.C,,H,,O, requires C = 58.89 ; H = 6.7 per cent. Equivalent = 316.The compound was readily soluble in the nsual organic solvents,and in sodium carbonate solution, less so in sodium hydrogencarbonate solution. It gave a Seep blue colour with ferric chloride.The following table shows the results obtained by a comparison ofthe melting point of the compound with those of the productsprepared according to the directions respectively of Blaise and ofPechmann, Bausr, and Obermiller :(1) The above product ........Mixture of (1) and (3) ...............78.5"(2) Blaise's ,, ........ 80-80*5" (Rlaise found 78-79")(3) Pechmann's ,, ......... 78.5-79" (Pechmann found 81-82")76-79.5",, ,, (1) ,, (3) ...............76-77",, ,, (2) ,, (3) ............. 78-79".The three products are therefore identical, and are to be regardedas ethyl 2 : 6-dicarbethosyd2 :hyclohexadier5-ol-l-acetate (see theo-retical part).The copper salt was prepared by shaking an ethereal solution of* Ethyl glntaconate was obtained according to the diredions of Blaise (BztZl. Soc.d i m . , 1903, [iv], 29, 1012) by the elimination of the elements of water from ethyl/3-hydroxyglutarate, the reduction product of acetonedicarboxylic acid. Withregard to the preparation of the latter acid, i t may he of value to emphasise thenecessity of proceeding with its isolation immediately carbon dioxide can bedetected in the gases evolved from the jnteraction of fuming sulphuric acid andcitric acid.This was found to be the case a t a very early stage of the heatingrecommended by Pechmann (Bey., 1884,17, 2543) after the preliminary rcaction hastaken place without the application of external heat.-I This is in agreement with the experience of Fecht (Zoc. cit.)THE CONDENSATION OF ETHYL GLUTACONATE. 289the ester with aqueous copper acetate solution, and remained asa brown powder after the green ethereal solution thus obtainedhad been evaporated. It crystallised from alcohol in flat needles,melting at 145O. The salt was insoluble in water or cold alcohol,moderately so in hot alcohol, and readily so in ether, chloroform,or benzene :0.4668 gave 0.0502 CuO. Cu =8.97.(C?,,H2,0,),Cu requires Cu = 8-91 per cent.The phenylhydrazide resulted when the ester (5 grams) washeated a t looo for fifteen minutes with phenylhydrazine (2 g r a b )and glacial acetic acid (12 drops).The mixture solidified wheni t was stirred with alcohol, and after crystallisation from thissolvent, truncated hexagonal pyramids were obtained, which melteda t 130° (Blaise gives 126-127O).Blaise experienced difficulty in the analysis of this compoundowing to the formation of carbon monoxide, but we were unableto detect this gas by the use of palladium chloride solution in thecourse of the .analyses quoted below, and no diminution in thevolume of nitrogen obtained was observed when it was left incontact with cuprous chloride solution :0.1612 gave 0.3746 CO, and 0.1013 H,O.0.2859The compound was insoluble in alkali, and underwent oxidationwhen its alcoholic solution was warmed with yellow mercuric oxide,a red azo-compound being produced.This behaviour shows thecompound to be a hydrazide and not a hydrazone (compareDieckmann, AnnaZert, 1901, 317, 60).C = 63.37 ; H = 5-66.0.1540 ,, 0.3605 GO2 ,, 0.0924 H20. C=63*85; H=5.94.,, 17.2 C.C. N2 a t 17O and 735 mm. N=6*88.C22H280sN2 requires C = 63.46 ; H = 5.98 ; N T= 6-73 per cent.Tnteraction of Ethyl 2 : 6-Dicarbethoxy-h2 :6-cyclo.hexadie~5-ol-l-acetate and Nydratine Hydrate.The addition of hydrazine hydrate (5 grams) to an alcoholicsolution of the ester (5 grams) caused the gradual separation of awhite compound a t the ordinary temperature, and the reaction wascomplete after a very short time a t looo.After being dried onporous earthenware, the product melted sharply a t 205-206O :0.2049 gave 0.3020 CO, and 0.1143 H,O. C =40.30; H = 6.24.0.0868C10H1403N6,N2H4 requires C = 40.27 ; H = 6-04 ; N = 37.58 per cent.The product was therefore a hydrazine salt of a pyrazolonederivative, C,,H,,O,N,. It was insoluble in alcohol, benzene, or,, 29.2 C.C. N, a t 19O and 729 mm. N=37-79290 THE CONDENSATION OF ETHYL GLUTSCONATE.chloroform, but readily soluble in water, alkali, or dilute acid. Itlost hydrazine only slowly a t looo, being converted into the yellowpyrazolone derivative. A yellowish-white silver salt separated whensilver nitrate was added t o an aqueous solution of the compound,rendered just acid with nitric acid.It was, however, unstable,and gradually underwent decomposition and reduction to a violet,and ultimately t o a black mass.The copper salt, prepared in a similar manner by the use ofcopper acetate, was sufficiently stable t o permit of its analysis:0.3924 gave 0.1164 CuO. Cu=23*7.C,oH,,03N,Cu requires Cu = 23.9 per cent.Hydro1ysi.s of Ethyl 2 : 6-Dicarbethoxy-b2 ~~-cyclohemdien-5-oJ-l-acetate.(1) A specimen of the ester, prepared according to Blaise’sdirections, was treated with potassium hydroxide under the con-ditions described by him. The diethyl ester thus obtained melteda t 110-1 12O (Pechmann gives 11 2-1 1 3 O , Blaise 98-99O). Themonoethyl ester melted and decomposed a t 167O (Pechmann gives157O, Blaise 1 7 8 O ) .(2) A specimen of the ester, prepared according to Pechmann’sdirections, was hydrolysed with 10 per cent. sulphuric acid underthe conditions described by Blaise. The product melted a t 66O,and agreed in its properties with the account given by Blaise.Further, it was oxidised by bromine in chloroform solution.O x i h t i o n of Ethyl 2 ; 6-Dicar6 ethtoxy-A2 ~~-cyclohexudien-5-ol-l-acetate.The ester, prepared according to the directions of Blaise, wastreated with bromine in chloroform solution (Pechmann, Bauer,and Obermiller, Zoc. cit.), whereupon ethyl I-hydroxy-2 : 4-dicarb-ethoxyphenylacetate (m. p. !30-81°) was obtained. It furnished atribasic acid (m. p. 249-250O) on hydrolysis (Pechmann gives250-255O) (Found, C = 49-95 ; H = 3-40. C,,H,07 requires C = 50.00;H=3*33 per cent.).THE UNIVERSITY,SHKFFIELD