RING-CHAIN TAUTOMERISM. PART N. 113 X.IV.-Ring-chain Tautomerism. Part I V. The Xff ect of the Methyl Ethyl Grouping on the Carbon T'etrahedral Angle. By BALBIR SINGH and JOCELYN FIELD THORPE. THE recent experiments of Deshapande and Thorpe (T. 1922 121, 1430) on pp-diethylglutaric acid have shown that in accordance with the general hypothesis regarding the influence of groups on t8he depression of the tetrahedral angle the effect produced by the gem-diethyl grouping lies approximately midway between those caused by the gemdimethyl group and by the six-carbon ring. On the other hand attention was directed to the very marked differences in chemical behaviour which accompanied the com-paratively small alteration in angle and particularly as regards the stability of the open-chain or ring system which was the outcome of the change.Thus whereas the keto-acid (I) is the stable form in the dimethyl series the ring hydroxy-acid (11) is the sta,ble structure in the six-ring series and as a matter of fact so much is this the case that it has not been found possible to prepare the ring hydroxy-acid (111) or the keto-acid (IV) in spite of the fact that the (111.) CH3>C<~(0H)*C02H ~ ~ 2 < c H 2 * ~ H 2 > ~ < ~ o * ~ o ' Z H (IV.) CH CH.CO,H CH,*CH CH,*CO,€I CHBr*CO,Et >C<CHBr*CO,Et (v.) starting point wits in each case the dibromo-ester (type V) or the corresponding bromo-lactone. These results were however, brought into line by the discovery by Deshapande and Thorpe that the keto-acid (VI) and the hydroxy-ring acid (VII) in the diethy 114 SINGH AND THORPB: series were tautomeric.That is to say they were mutually inter-convertible in strong potash solution; a discovery which again emphasised the wide chemical differences produced by a slight angle alteration because the keto-acid (I) and the hydroxy-ring acid (11) are unaltered by alkali. There are moreover other points of difference which reveal themselves when the experimental data derived from the three series (dimethyl diethyl and cyczohexane) are compared. For example when the hydrolysis of the dibromo-ester of type V is carried out with ethyl alcoholic potash the products in the cases of the diethyl and cyclohexane series are the same as with aqueous potassium hydroxide whereas in the dimethyl series the sole product is the ethoxy-derivative (VIII) and no trace of the keto-acid is produced.This remarkable reaction was considered by Perkin and Thorpe (T. 1901 79 729) to be due to the addition of alcohol t o the unsaturated compound (IX). It is evident that this fundamental difference between the dimethyl and diethyl series must rest on the small difference of angle in the two series and in order to study the effect produced by diminishing this difference we have carried out a series of comparative experiments in the methyl ethyl series. pB-Methylethylglutaric acid is readily prepared from methyl ethyl ketone by the Guareschi method. The anhydride dibrominates normally but the dibromo-ester (X) loses ethyl bromide when distilled and passes into the bromo-lactonic ester (XI) (compare Beasley Ingold and Thorpe T.1915 107 1080). (X.) CH3>C<CHBr*C02Et -j. cH3>c<cH(co2Et)*p (XI.) C2H5 CHBr*CO,Et C,H5 CHBr- co The dibromo-acid (XII) was isolated in cis- and trans-forms both of which gave the same hydroxylactonic acid (XIII) with sodium carbonate. Like the corresponding lactonic acid in the diethyl (=I.) CH3>C< CHBr*C02H CH3>c<CH(C02H)*? (XIII.) C2H5 CHBr*CO,H C2H5 CH(0H)-CO series the acid (XIII) is unaltered by strong aqueous caustic alkali, and there is thus no tautomerism between the hydroxylactone-hydroxy-ring acid and keto-acid such as was shown to be the case with Balbiano's acid (Kon Stevenson and Thorpe T. 1922 121, 650). Indeed the recurrence of such tautornerism was of course, not to be expected although in this instance the proof of structur RING-CHAIN TAUTOMERISM.PART IV. 115 was supplied by the formation of a well-defined acetyl derivative (XIV). When the dibromo-ester (X) is hydrolysed by means of ethyl alcoholic potash it behaves like the corresponding dibromo-ester in the dimethyl series and yields the ethoxy-ring acid (XV). The stability of this substance is remarkable because it is unaltered by strong aqueous alkali even on prolonged boiling and is un-affected by concentrated hydrochloric acid. (XIV.) CH3>C<CH(C02H)*Q CH3>C< C(OEt)*CO,H I (XV.) (XVI.) CH,\ WMe)*C02I3: C2H5 CH(OAc)-CO C2H5 CH*CO,H C2H,/ c<LH*cO,H The ethoxy-acid (XV) is partly converted into the ring-acid (XXII) with strong hydriodic acid and gives the correct " ethoxy-value " by the Zeisel method.It is however with the correspond-ing methoxy-derivative (XVI) that we have been able to carry out experiments which have enabled us to co-ordinate the formation of these apparently abnormal substances wit'h the other reactions of the series and to show that their formation and properties are in every way consistent with our general hypothesis. This remarkable substance can be isolated in well-defined cis-and trans-forms the trans-modification being converted into the cis- through the anhydride of the latter. Both isomerides pass on treatment with hydriodic acid into the keto-acid (XVII) which can be characterised by the formation of a well-defined quinoxaline derivative (with o-phenylenediamine) and by its conversion on distillation into the anhydride of methylethylsuccinic acid (XVIII) (compare Balbiano's acid T.1922 121 650). Here then we have (XVII.) CH3>C<CO0(W3 CH3>c<CO-? (XVIII.) C2H5 CH2*C0,H C2H5 CH2*C0 (XIX.) CH3>C<~(0H)*c02H C2H5 CH*CO,H the mechanism of the change revealed because it will be noticed that the action of hydriodic acid is not a reducing action but merely one by which the group OCH is converted into OH in the usual manner. The hydroxy-acid (XIX) is however unstable and passes a t once into the open-chain keto-acid (XVII) which is stable to 64 per cent. aqueous potassium hydroxide. These experiments therefore, afford complete confirmation of our contention that the alteration of the tetrahedral angle which causes the keto-acid and hydroxy-ring acid to be tautomeric in the case of the gem-diethyl compomd, causes also the hydroxy-ring acid to be the only stable form in th 116 SINGH AND THORPE: cyclohexane seiies and the keto-acid the only stable form in the gem-dimethyl series.The reasons why the alkyloxy-derivatives are formed a t all in the instances given are also quite clear but as these will be dealt with fully in a future paper and are bound up with the chemistry of the cyclopropene structures to be mentioned there it is unnecessary to give them in this communication. It is sufficient here to mention that the isolation of these abnormal alkyloxy-derivatives which was carried out twenty-two years ago and which the later work of Feist on the cyclopropene acid (XX) rendered still more dificult to explain has now proved to be the key to the whole problem of reversible and non-reversible isomerism in this series.Feist's (=.I CII,CH<I C*C02H I (JH <g*Co,H (XXI.) C*CO,H C*CO,H acid and still more recently the acid (XXI) (Farmer and Ingold, T. 1921 119 2001) did not tend to add on alcohol because they constituted stable glutaconic systems each cont'aining the necessary mobile hydrogen atom. On the other hand the acids of the pp-gem-series cannot contain the hydrogen atom and in consequence are incapable of isolation in the unsaturated form and even when the unsaturated bond is removed by addition the ring remains unstable unless the tetrahedral angle has been sufficiently deflected. It is scarcely necessary to point out that the addition of alcohol leads to a stable alkyloxy-derivative and the addition of water to either the keto-acid or hydroxy-ring acid whichever is the more stable because the presence of the alkyloxy-group does not supply the necessary hydrogen atom for ring rupture.For the rest the ring a,cid (XXII) occurs in well-defined cis- and trans-forms and shows all the usual reactions of acids of this type. As regards stability they do not undergo fission with mineral acids, and are therefore more stable than the caronic acids which are completely decomposed under these conditions into terebic acid (compare Perkin and Thorpe T. 1899 75 48). This is therefore, another indication of stability caused by the alteration of the tetrahedral angle. Again the lactonic acid (XIII) is evidently the lactonic acid of the trans-dihydroxy-acid because the corresponding cis-dihydroxy-acid (XXIII) is stable and gives a characteristic silver sal RING-CIEAM TAUTOMERISM.PART IV. 117 E X P E R I M E N T A L . Methylethylglutaric acid was prepared by the condensation of methyl ethyl ketone with ethyl cyanoacetate in the presence of alcoholic ammonia (Guareschi's general method as modified by Kon T. 1921 119 818). It was obtained in quite a satisfactory yield. The anhydride was prepared by boiling the acid with acetic anhydride and distilling the product under diminished pressure. The anhydride sets to a crystalline mass on cooling and melts above 25". It reacts with aniline in benzene solution in the cold to form the unilic acid which was obtained in a crystalline con-dition (compare T.1911 99 440) and when recrystallised from a mixture of chloroform and petroleum melted at 105" (Found: C = 67.3; H = 7.5. C1,H,,03N requires C = 67.4; H = 7.6 per cent .). Dibromination of pp-~~ethylelhylglutaric Acid. Dibromination was effected by converting the acid or its anhydride into the acid chloride and then adding bromine. In the case of the acid the following coiiditions were observed : Eighty-seven grams of the acid were mixed with 208 grams of phosphorus pentachloride and the mixture was gently wrrmed on the water-bath until a clear solution was obtained. To this 67 C.C. of dry bromine were added in four instalments and the mixture was heated on the water-bath until free from halogen.After cooling, the product was divided; part being poured into anhydrous formic acid and part into absolute alcohol. trans- ad- Di brorno- p p-meth ylethylglutaric Acid (XII) .-The formic acid solution was gently warmed on the water-bath for two hours and then left in a crystallising dish at the ordinary temperature for slow evaporation. The first crop of the crystals that separated consisted almost exclusively of the trans-isomeride. Drained on porous plate washed with dry benzene and twice crystallised from a mixture of chloroform and light petroleum (b. p. 100-120") it melted sharply and decomposed at 173" (Found Br = 48.2. C8HI2O0Br2 requires Br = 48.1 per cent.). A further quantity was isolated by treating the residual semi-solid mass obtained after the formic acid had completely evaporated with dry benzene.The insoluble granular portion was collected by the aid of the pump and purified in the same way as the first crop. cis-ud-Dibromo- p p-methylethylglutn ric Acid .-The cis-isomcride was obtained from the benzene-soluble portion and purified by repeated crystallisations alternately from benzene and a mixture o 118 SINGH AND THORPE : chloroform and light petroleum. The pure acid melts at 146" (Found Br = 48.0. Ethyl oru'-Dibromo- pp-methylethylglutarate (X) .-The crude product of dibromination poured into absolute alcohol was kept over-night . It was then diluted with water and extracted with ether. The ethereal solution was thoroughly washed with dilute sodium carbonate solution and dried.On evaporating the crude dibromo-ester was obtained as a pale coloured oil which could not be dis-tilled under diminished pressure without decomposition. From one sample a constant-boiling fraction was obtained (194-196"/31 mm.), and it proved to be the bromo-lactonic ester of methylethylgluturic acid (XI) (Found Br = 28.5. C,,H,,O,Br requires Br = 28.6 per cent .). In the subsequent experiments we therefore used the crude dibromo-ester. Its analysis gave the following result Found : Br = 37.9. C,,H,,,O,Br requires Br = 41.2 per cent. Acid Ester.-The sodium carbonate washings of the ethereal solution of the neutral dibromo-ester contained just traces of the acid ester. If however thionyl chloride be substituted for phosphorus pentachloride a considerable proportion of the acid ester is in-variably produced.It does not distil without decomposition. That it is mainly a monobromo-acid ester is proved from the products obtained on hydrolysing it (p. 121). C,H,,O,Br requires Br = 48.1 per cent.). trans-Lactonic Acid of ua'- Dih ydroxy- p P-meth yleth ylglutaric Acid (XIII). Twenty grams of the trans-dibromo-acid (XII) were dissolved in 200 C.C. of 2N-solution of sodium carbonate and the solution was boiled for two hours. When cold it was acidified with hydro-chloric acid and extracted with ether ; the ethereal solution yielding a gum on evaporation which went completely solid after keeping. Recrystallised from benzene it melted at 116" (Found C = 50*9; H = 6.3. C,H1,05 requires C = 51.0; H = 6.3 per cent.).It is moderately soluble in all the usual solvents except light petroleum in which it is practically insoluble. The trans-hydroxy-lactonic acid is a remarkably stable compound. It was recovered unchanged after treatment with boiling 64 per cent. potassium hydroxide solution. Still higher concentrations of alkali however brought about deep-seated decomposition oxalic acid being the only substance that was identified in the products of decomposition. The cis-dibromo-acid yielded the same trans-lactonic acid under similar conditions of hydrolysis RMG-CHAIN TAUTQNERISM. PART IV. 119 The silver salt is extremely soluble in water and could not be collected in quantity sufficient for analysis. The acetyl derivative (XIV) was obtained by boiling the acid with a sufficient quantity of acetyl chloride for fifteen minutes.It separated from benzene in small needles melting at 121". It is readily soluble in ether chloroform or acetone and moderately soluble in benzene (Found C = 52.1 ; H = 5.6. CloH1,O, requires C = 52.1; H = 6.0 per cent.). 1 - Ethox y -3-meth yl- 3- eth ylcy clopropane- 1 2 -d ica rboxylic Acid (XV) * Sixty grams of the dibromo-ester (X) were mixed with an alcoholic potassium hydroxide solution containing 60 grams of the hydroxide, 36 C.C. of water and 240 C.O. of ethyl alcahol and boiled under reflux for three hours. The solution was then freed from alcohol, acidified and extracted with ether. The syrupy residue obtained on evaporation could not be induced to crystalhe. However on treatment with ammonium hydroxide solution (d 0.880) a crystalline ammonium salt of the ethoxy-acid was isolated which was further purified by dissolving in alcohol and precipitating by the addition of dry ether.The purified ammonium salt was dissolved in dilute hydrochlorio acid and extracted with ether when the ethereal solution deposited crystals of the ethoxy-acid on evaporation. Recrystallised from benzene it melted at 150" [Found C = 55.0; H = 7.6 ; OEt = 20.3. C,H,,Q,(OEt) requires C = 55.5 ; H = 7.4 ; OEt = 20.8 per cent. 3. It was recovered unchanged (apart from the production of a little tarry matter) after having been boiled for several hours with hydrochloric acid. A treatment with strong alkali solution likewise did not effect any change However, gentle heating with 80 per cent.sulphuric acid decomposed it into aa-methylethylsuccinic acid melting at 103" with the evolution of carbon monoxide (Found C = 52.6 ; H = 7.5. Calc. C = 52.5 ; H = 7.5 per cent.). By boiling the ethoxy-acid with hydriodic acid a gummy mass was obtained from which a small quantity of trans-3-methyl-3-ethylcyclopropane- 1 2-dicarboxylic acid (XXII) was isolated. The ethoxy-acid is very stable. Action of Xtrong Methyl Alcoholic Potash on Ethyl aa'-Dibrommnethyl-ethylglutarate Formation of trans- and cis-l-Methoxy-3-methyE-3-ethylcyclopropane-1 2-dicarboxylic Acid and ad-Dihydroxy- pp-methylethylgluturic Acid. To a boiling solution of 202 grams of potassium hydroxide in 600 C.C. of methyl alcohol were added 129 grams of the neutra 120 SINGH AND THORPE : dibromo-ester (X) as rapidly as possible and the heating was continued for twenty minutes.After evaporating off the alcohol, the residue was dissolved in water acidified and extracted with ether and the ethereal solution dried over calcium chloride. ad-Dihydroxy- PP-methylethylglutaric Acid (XXIII).-The calcium chloride used for drying the ethereal solution was found to have changed partly into a calcium salt of an organic acid. The whole of the calcium chloride mixed with the organic salt was therefore dissolved in dilute hydrochloric acid and extracted with ether. The ethereal solution on slow evaporation deposited beautiful long needles of ad-dihydroxy- PP-methylethylglutaric acid. Recrystal-lised from chloroform it melted at 80" (Found C = 46.5 ; H = 6.6.C8H,06 requires C = 46.6; H = 6.7 per cent.). The silver salt was prepared in the usual way. It is a crystalline, white powder (Found Ag = 51.6. C8Hl,06Ag requires Ag = 51.4 per cent. ). Acid (XVI).-The dried ethereal solution remaining after separation of the calcium salts was evaporated and a gum was obtained which completely solidified on keeping. From this by fractional crystal-lisation from chloroform the cis- and the trans-isomerides of the methoxy-acid were separated. The trans-acid when pure melts at 149" and crystallises well from a mixture of acetone and light petroleum. It is less soluble in chloroform than its isomeride (Found C = 53.6; H = 6.7. C,Hl,O requires C = 53.4; H = 6.9 per cent.). It is freely soluble in chloroform, acetone and ethyl acetate (Found C = 53.4; H = 6.7.CgH140, requires C = 53.4 ; H = 6.9 per cent.). The Anhydride.-The cis-acid on boiling with acetyl chloride is readily converted into the anhydride from which the original acid can be regenerated by boiling it with water. The anhydride is a colourless mobile liquid boiling at 144"/7 mm. (Found C = 58.2 ; H = 6.0. CgH1,04 requires C = 58.6 ; H = 6.5 per ccnt.). The trans-acid when heated with acetic anhydride in a sealed tube is changed into the anhydride of the cis-acid from which the cis-acid can be produced by hydration. a-Keto- PP-methylethylglutaric Acid (XVII).-This was obtained by hydrolysing the methoxy-acid (XVI) with hydriodic acid. A mixture of 32 grams of the methoxy-acid and 160 C.C.of hydriodic acid (d 1.7) was gently heated on the oil-bath. After the temperature had been maintained at 80" for some time during which period the methyl iodide distilled off it was raised to the boiling point of thc solution the whole operation taking one and a half hours. When 1 -Hethox y-3-methyl- 3-eth ylcyclopropane- 1 2-dicarbox ylic The cis-acid melts at 141' RING-CHAIN TAUTOMERISM. PART IV. 121 cold the solution was treated with just enough sulphurous acid to reduce free iodine and was extracted with ether the ethereal solution yielding the fluid keto-acid on evaporation. This could not be induced to crystallise. It was however characterised by the formation of a well-defined quinoxaline derivative which was readily formed when the keto-acid was intimately mixed with o-phenylenediamine in acetic acid solution and gently warmed.Recrystallised from acetone diluted with water the quimxuline melted a t 216" (Found C == 64.6; H = 6.2. C14€I1603N2 requires C = 64.6; H = 6.1 per cent.). The keto-acid when distilled under diminished pressure is converted into the anhydride of methylethylsuccinic acid (XVIII), from which methylethylsuccinic acid was obtained and identified. Hydrolysis of the Monobromo-acid Ester. The monobromo-acid ester was obtained as a by-product in the experiment of dibromination of the methylethylglutaric acid and was hydrolysed with a strong aqueous solution of potassium hydroxide. trans-3-iklethyl-3- ethylcyclopropane- 1 2-dicarboxylic Acid (XXII) . -A solution containing 60 grams of potassium hydroxide in 50 C.C.of water was heated to boiling. Nineteen grams of the mono-bromo-acid ester were then cautiously added as rapidly as possible. A vigorous reaction ensued which gradually subsided. When cold, the whole was dissolved in water and acidified with hydrochloric acid when crystals of the trans-acid separated on cooling. The whole of the trans-acid was thus separated by filtration. Re-crystallised from ethyl acetate it melted a t 221" (Found C = 65.0 ; H = 6.8. C,H,,O requires C = 55.8; H = 6.9 per cent.). The silver salt is a white powder (Found Ag = 55.6. C,H,O,Ag, requires Ag = 55.9 per cent.). The dianilide was prepared by heating the acid with excess of aniline a t 200" for forty-five minutes. The product separated from dilute alcohol in light scales melting a t 291" (Found C = 75.0 ; H = 7.1.Acid. - The filtrate from the trans-acid was extracted with ether the ethereal solu-tion yielding a semi-solid mass on evaporation. This was spread on a porous tile drained and washed with a small quantity of water, which caused the cis-acid to become solid. The cis-acid was also prepared by heating the trans-acid with acetic anhydride in a sealed tube and treating the product with water. It usually melts a t 163-165" but repeated crystallisation from chloroform finally raises its melting point to 180" (with decomposition) (Pound C = 55.5; H = 6.8 C,,,H,,O,N requires C = 74.5; H = 6.8 per cent.). cis-3-Methyl-3-ethylcyclopropane- 1 2-dicarboxylic C,HI2O4 requires C = 55%; H = 6.9 per cent.) 122 BAKER AND INGOLD THE FORMATION The anhydride of the cis-acid was prepared by heating the acid above its melting point until the evolution of water vapour had ceased. It was characterised by conversion into the an& acid, which was obtained by mixing a benzene solution of the anhydride with a solution of aniline in the same solvent. Crystals separated on agitating the solution with a glass rod. Recrystallised from dilute alcohol it melted and decomposed at 193" (Found C = 68.4 ; H = 6% C,,H,,O,N requires C = 68.0; H = 6.8 per cent.). We are indebted to the Chemical Society for a grant which has IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY, partly defrayed the cost of the materials used in the investigation. [Received December 16th 1922.1 SOUTH KENSINQTON S.W